• OpenAI's o1 model was hailed as a breakthrough in September.
• By November, a Chinese AI lab had released a similar model called DeepSeek.
• On Thursday, Google came out with a challenger called Gemini 2.0 Flash Thinking.

In September, OpenAI unveiled a radically new type of AI model called o1. In a matter of months, rivals introduced similar offerings.

On Thursday, Google released Gemini 2.0 Flash Thinking, which uses reasoning techniques that look a lot like o1.

Even before that, in November, a Chinese company announced DeepSeek, an AI model that breaks challenging questions down into more manageable tasks like OpenAI's o1 does.

This is the latest example of a crowded AI frontier where pricey innovations are swiftly matched, making it harder to stand out.

"It's amazing how quickly AI model improvements get commoditized," Rahul Sonwalkar, CEO of the startup Julius AI, said. "Companies spend massive amounts building these new models, and within a few months they become a commodity."

The proliferation of multiple AI models with similar capabilities could make it difficult to justify charging high prices to use these tools. The price of accessing AI models has indeed plunged in the past year or so.

That, in turn, could raise questions about whether it's worth spending hundreds of millions of dollars, or even billions, to build the next top AI model.

September is a lifetime ago in the AI industry

When OpenAI previewed its o1 model in September, the product was hailed as a breakthrough. It uses a new approach called inference-time compute to answer more challenging questions.

It does this by slicing queries into more digestible tasks and turning each of these stages into a new prompt that the model tackles. Each step requires running a new request, which is known as the inference stage in AI.

This produces a chain of thought or chain of reasoning in which each part of the problem is answered, and the model doesn't move on to the next stage until it ultimately comes up with a full response.

The model can even backtrack and check its prior steps and correct errors, or try solutions and fail before trying something else. This is akin to how humans spend longer working through complex tasks.

DeepSeek rises

In a mere two months, o1 had a rival. On November 20, a Chinese AI company released DeepSeek.

"They were probably the first ones to reproduce o1," said Charlie Snell, an AI researcher at UC Berkeley who coauthored a Google DeepMind paper this year on inference-time compute.

He's tried DeepSeek's AI model and says it performs well on complex math problems that must be solved by thinking for longer and in stages. 

He noted that in DeepSeek's DeepThink mode, the model shows users every step of its thought process. With o1, these intermediate steps are hidden from users. 

"I've asked people at OpenAI what they think of it," Snell told BI. "They say it looks like the same thing, but they don't how DeepSeek did this so fast."

OpenAI didn't respond to a request for comment. On Friday, the startup previewed an o1 successor, called o3. Francois Chollet, a respected AI expert, called the update a "significant breakthrough."

Andrej Karpathy, an OpenAI cofounder, praised Google's new "Thinking" model for the same reasoning feature.

"The prominent and pleasant surprise here is that unlike o1 the reasoning traces of the model are shown," he wrote on X. "As a user I personally really like this because the reasoning itself is interesting to see and read — the models actively think through different possibilities, ideas, debate themselves, etc., it's part of the value add."

A DeepSeek demo

Snell shared a multistep math problem with Business Insider, which we used to test DeepSeek for ourselves: 

"Find a sequence of +, -, /, * which can be applied to the numbers 7, 3, 11, 5 to get to 24, where each of the given numbers is used exactly once."

BI put that prompt in DeepSeek's chat window on its website. The model responded initially by laying out the challenge ahead.

"Alright, so I've got this problem here: I need to use the numbers 7, 3, 11, and 5, and combine them with the operations of addition, subtraction, multiplication, and division, using each number exactly once, to get to 24," it replied. "At first glance, this seems a bit tricky, but I think with some systematic thinking, I can figure it out."

It then proceeded through multiple steps over roughly 16 pages of discussion that included mathematical calculations and equations. The model sometimes got it wrong, but it spotted this and didn't give up. Instead, it swiftly moved on to another possible solution. 

"Almost got close there with 33 / 7 * 5 ≈ 23.57, but not quite 24. Maybe I need to try a different approach," it wrote at one point. 

After a few minutes, it found the correct solution. 

"You can see it try different ideas and backtrack," Snell said in an interview on Wednesday. He highlighted this part of DeepSeek's chain of thought as particularly noteworthy:

"This is getting really time-consuming. Maybe I need to consider a different strategy," the AI model wrote. "Instead of combining two numbers at a time, perhaps I should look for a way to group them differently or use operations in a nested manner."

Then Google appears

Snell said other companies are likely working on AI models that use the same inference-time compute approach as OpenAI.

"DeepSeek does this already, so I assume others are working on this," he added on Wednesday.

The following day, Google released Gemini 2.0 Flash Thinking. Like DeepSeek, this new model shows users each step of its thought process while tackling problems. 

Jeff Dean, a Google AI veteran, shared a demo on X that showed this new model solving a physics problem and explained its reasoning steps. 

"This model is trained to use thoughts to strengthen its reasoning," Dean wrote. "We see promising results when we increase inference time computation!"

• Rohan Doctor was a managing director at Goldman Sachs when he founded Louisa AI.
• The startup uses AI to feed deal ideas and networking prompts to bankers and investors.
• Here's why he wants to bring the dealmaking playbook to startups.

Cold call after cold call, Rohan Doctor wasn't getting as far as he would've hoped.

The former Goldman Sachs managing director had emailed a list of digital strategy execs at banks and private equity firms to try to sell them on his startup, Louisa AI. But he only got a handful of replies back.

Two years since its launch, Louisa AI has secured about a dozen clients, including some of the biggest names in corporate America. They include Goldman Sachs, VC firm Insight Partners, and, more recently, one of the biggest AI chipmakers and a top consulting firm. But he didn't secure those contracts from cold outreach. He used his own startup's technology, which proactively prompts deal ideas based on people's personal and professional connections, to get in through the front door.

Now, Doctor wants to bring his dealmaking playbook to other startups ahead of an anticipated M&A boom.

"If we're able to close more deals through warm relationships this way, then other startups can, too," he said.

The near-term outlook for M&A activity has gotten brighter, with lower interest rates reducing the cost of borrowing. Wall Street execs are optimistic that Trump's return to the White House, and any business-friendly regulations that may come with it, will be a tailwind for dealmaking. Also, companies resetting their valuations could spur more transactions to close as price expectations align between buyers and sellers.

Meanwhile, in Silicon Valley, VCs and founders are hopeful about the anticipated looser environment, which could boost tech building and dealmaking. VCs, which rely on selling startups in M&A deals for many of their returns, have been dampened by the Federal Trade Commission's antitrust stance on M&A.

Louisa AI was built to suggest potential deals based on the data it's exposed to. It ingests information about who and what employees know by plugging into company CRMs, messaging platforms like Slack and Symphony, and email providers. Since spinning out of Goldman Sachs in 2023, Louisa AI has raised $5 million in seed funding. It suggests about $1 billion in deal values per quarter, Doctor said.

It also highlights mutual connections to establish a warm introduction, which can make all the difference in the multi-billion investment banking industry built on relationships. While running the bank-solutions group at Goldman Sachs, Rohan Doctor used his network to close transactions worth tens of millions of dollars. As a startup founder, it's been a different story.

"I've tried the cold outreach and just emailing," Doctor said, adding that the startup stopped doing that after it didn't yield good results. What has worked for Doctor is realizing he knows someone who knows someone.

Louisa AI scored the chipmaker contract after the AI flagged that one of Doctor's staff used to work for someone who now worked at the chip manufacturer. With the consultancy, one of Louisa AI's investors connected Doctor with the consulting firm they used to work for. He declined to name these firms due to non-disclosure agreements.

"Everything needs to be warm when it comes to big companies doing big things with other people. It has to rely on trust," he said.

• Verily, an Alphabet spinoff, plans to raise money and focus its strategy on healthcare AI in 2025.
• It plans to sell tech tools that other companies can use to build AI models and apps.
• The changes are underway as Verily separates itself from Alphabet and looks to mature as a company.

Verily Life Sciences plans to reorient its strategy around AI in 2025, just as it marks its 10th anniversary as a company under Alphabet.

The unit, which uses technology and data to improve healthcare, is looking to mature. As of January, it will have separated from many of Google's internal systems in an attempt to stand independently. Simultaneously, it's refocusing its strategy around AI, according to two employees with knowledge of the matter, who asked to remain anonymous to discuss confidential information.

This new strategy would primarily involve other healthcare companies using Verily's tech infrastructure to develop AI models and apps, resulting from a multi-year effort across teams. It ultimately aims to become companies' one-stop-shop for tech needs like training AI for drug discovery and building apps for health coaching.

The unit is also looking to raise another round of capital in the next year, the two people familiar with the matter said. The company's last investment was a $1 billion round led by Alphabet in 2022. Alphabet will likely lead the round again, although leadership could also bid for outside capital as Verily tries to become "increasingly independent," one source said.

The question for next year is whether Verily can finally start turning long-gestating ideas into profits. One of the people said Verily still generates the most revenue selling stop-loss insurance to employers, which is a far cry from the higher-margin business it's aiming for. The Wall Street Journal reported last year that this business, called Granular Insurance, was Verily's most lucrative.

Verily has been criticized in the past for having a rudderless strategy. It's entertained bets on topics as diverse as editing mosquito populations and helping pharmaceutical companies run clinical trials.

In an email to Business Insider, a spokesperson for Verily declined to comment on non-public information. He confirmed the company's plans to provide tech infrastructure for third parties, designed to provide "precision health capabilities across research and care."

The AI strategy's origin story

Verily's idea to become a tech provider for other healthcare companies grew out of its own internal needs a few years ago when it decided to "re-platform" its various bets on a shared infrastructure, a source familiar with the matter said.

The multi-year effort is now coming to fruition, and Verily plans to sell the core technology it uses to health plans, providers, digital health startups, and life sciences companies.

The platform will include data storage and AI training. Companies could also use Verily's tech tools to spin up apps without having to code as much. For example, a digital health startup could use Verily's tools to build a coaching app with AI insights on weight loss.

"Large pharma companies, for example, look at the work we do and recognize that the data science applications or clinical research tools that they need to build themselves could be better if they were built using our platform," said Verily CEO Stephen Gillett in an interview with Fortune in November.

In that interview, Gillett said Verily's tech tools would include sophisticated AI capabilities for healthcare, data aggregation, privacy, and consent. One source said the company plans to start rolling them out in 2025.

Even as the leading AI models learn from the entirety of the internet, healthcare data remains largely private. Subsequently, Verily is betting that there's a growing need to further specialize the models for patient care and research. The upstart already does this work through its partnership with clients like the National Institutes of Health. Through a business line called Workbench, Verily hosts massive datasets for the NIH, complete with analysis tools.

Verily hasn't dropped its ambitions to grow its own healthcare business. In 2026, it plans to relaunch a diabetes and hypertension-focused app, Lightpath, broadly for health plans and employers — this time with AI coaches supplementing human ones. Verily also intends to expand Lightpath to more health conditions.

Verily's reshuffling

Verily spun out of Google's moonshot group in 2015 and remained part of Alphabet's collection of long-term businesses, sometimes called "other bets." Under its then-CEO Andy Conrad, the unit explored a menagerie of ideas from surgical robots to wearables. Several of these projects — glucose-monitoring contact lenses, for instance — haven't panned out.

Shortly after Gillett replaced Conrad as CEO in 2023, he announced the company would lay off 15% of its workforce and "advance fewer initiatives with greater resources."

Since then, Verily has pruned projects and teams to save costs and sharpen its focus. Dr. Amy Abernethy, Verily's former chief medical officer who joined the company in 2021, focused on aiding clinical research before departing late last year.

Verily's shift to AI, meanwhile, seems to have coincided with the hiring of Myoung Cha and Bharat Rajagopal as the chief product and revenue officers, respectively, earlier this year.

Cutting ties with Google

Executing the AI strategy isn't the only challenge Verily's leadership faces in 2025.

Since 2021, the life science unit has been reducing its dependency on Google's internal systems and technology through an internal program known as Flywheel. BI previously reported that it set a December 16, 2024, deadline to cut many of these ties.

The separation involves Verily employees losing many of their cushy Google benefits, which has been a point of consternation for the group, the two people said.

Gillett remarked in a town hall meeting earlier this year that some employees may feel Verily is no longer the place for them after the separation, according to a person who heard the remarks.

• Generative AI tools have made it easier to create fake images, videos, and audio.
• That sparked concern that this busy election year would be disrupted by realistic disinformation.
• The barrage of AI deepfakes didn't happen. An AI researcher explains why and what's to come.

Oren Etzioni has studied artificial intelligence and worked on the technology for well over a decade, so when he saw the huge election cycle of 2024 coming, he got ready.

India, Indonesia, and the US were just some of the populous nations sending citizens to the ballot box. Generative AI had been unleashed upon the world about a year earlier, and there were major concerns about a potential wave of AI-powered disinformation disrupting the democratic process.

"We're going into the jungle without bug spray," Etzioni recalled thinking at the time.

He responded by starting TrueMedia.org, a nonprofit that uses AI-detection technologies to help people determine whether online videos, images, and audio are real or fake.

The group launched an early beta version of its service in April, so it was ready for a barrage of realistic AI deepfakes and other misleading online content.

In the end, the barrage never came.

"It really wasn't nearly as bad as we thought," Etzioni said. "That was good news, period."

He's still slightly mystified by this, although he has theories.

First, you don't need AI to lie during elections.

"Out-and-out lies and conspiracy theories were prevalent, but they weren't always accompanied by synthetic media," Etzioni said.

Second, he suspects that generative AI technology is not quite there yet, particularly when it comes to deepfake videos. 

"Some of the most egregious videos that are truly realistic — those are still pretty hard to create," Etzioni said. "There's another lap to go before people can generate what they want easily and have it look the way they want. Awareness of how to do this may not have penetrated the dark corners of the internet yet."

One thing he's sure of: High-end AI video-generation capabilities will come. This might happen during the next major election cycle or the one after that, but it's coming.

With that in mind, Etzioni shared learnings from TrueMedia's first go-round this year:

• Democracies are still not prepared for the worst-case scenario when it comes to AI deepfakes.
• There's no purely technical solution for this looming problem, and AI will need regulation. 
• Social media has an important role to play. 
• TrueMedia achieves roughly 90% accuracy, although people asked for more. It will be impossible to be 100% accurate, so there's room for human analysts.
• It's not always scalable to have humans at the end checking every decision, so humans only get involved in edge cases, such as when users question a decision made by TrueMedia's technology. 

The group plans to publish research on its AI deepfake detection efforts, and it's working on potential licensing deals. 

"There's a lot of interest in our AI models that have been tuned based on the flurry of uploads and deepfakes," Etzioni said. "We hope to license those to entities that are mission-oriented."

• Andrej Karpathy recently suggested AI could enhance e-book reading with interactive features.
• Amazon may already be thinking about this for its Kindle e-books.
• The company is looking for an applied scientist to improve the reading and publishing experience.

The AI pioneer and OpenAI cofounder Andrej Karpathy thinks AI can significantly improve how people read books. Amazon may already be thinking about how to do this for its Kindle e-books business.

In a series of posts on X this week, Karpathy proposed building an AI application that could read books together with humans, answering questions and generating discussion around the content. He said it would be a "huge hit" if Amazon or some other company built it.

A recent job post by Amazon suggests the tech giant may be doing just that.

Amazon is looking for a senior applied scientist for the "books content experience" team who can leverage "advances in AI to improve the reading experience for Kindle customers," the job post said.

The goal is "unlocking capabilities like analysis, enhancement, curation, moderation, translation, transformation and generation in Books based on Content structure, features, Intent, Synthesis and publisher details," it added.

The role will focus on not just the reading experience but also the broader publishing and distribution space. The Amazon team wants to "streamline the publishing lifecycle, improve digital reading, and empower book publishers through innovative AI tools and solutions to grow their business on Amazon," the job post said.

3 phases

Amazon identified three major phases of the book life cycle and thinks AI could improve each one.

• First up is the publishing part where books are created.
• Second is the reading experience where AI can help build new features and "representation" in books and drive higher reading "engagement."
• The third stage is "reporting" to help improve "sales & business growth," the job post said.

An Amazon spokesperson didn't immediately respond to a request for comment on Friday.

'I love this idea'

There seems to be huge demand for this type of service, based on the response to Karpathy's X post.

Stripe CEO Patrick Collison wrote under the post that it's "annoying" to have to build this AI feature on his own, adding that it would be "awesome when it's super streamlined."

Reddit's cofounder Alexis Ohanian wrote, "I love this idea."

Do you work at Amazon? Got a tip?

Contact the reporter, Eugene Kim, via the encrypted-messaging apps Signal or Telegram (+1-650-942-3061) or email ([email protected]). Reach out using a nonwork device. Check out Business Insider's source guide for other tips on sharing information securely.

• On Monday, OpenAI released Sora, an AI video generator, in hopes of helping creators.
• One such creative, Marques Brownlee, wants to know if his videos were used to train Sora.
• "We don't know if it's too late to opt out," Brownlee said in his review of Sora.

On Monday, OpenAI released its Sora video generator to the public.

CEO Sam Altman showed off Sora's capabilities as part of "Shipmas," OpenAI's term for the 12 days of product launches and demos it's doing ahead of the holidays. The AI tool still has some quirks, but it can make videos of up to 20 seconds from a few words of instruction.

During the launch, Altman pitched Sora as an assistant for creators and said that helping them was important to OpenAI.

"There's a new kind of co-creative dynamic that we're seeing emerge between early testers that we think points to something interesting about AI creative tools and how people will use them," he said.

One such early tester was Marques Brownlee, whose tech reviews have garnered roughly 20 million subscribers on YouTube. One could say this is the kind of creator that OpenAI envisions "empowering," to borrow execs' term from the livestream.

But in his Sora review, posted on Monday, Brownlee didn't sugarcoat his skepticism, especially about how the model was trained. Were his own videos used without his knowledge?

This is a mystery, and a controversial one. OpenAI hasn't said much about how Sora is trained, though experts believe the startup downloaded vast quantities of YouTube videos as part of the model's training data. There's no legal precedent for this practice, but Brownlee said that to him, the lack of transparency was sketchy.

"We don't know if it's too late to opt out," Brownlee said.

In an email, an OpenAI spokesperson said Sora was trained using proprietary stock footage and videos available in the public domain, without commenting on Business Insider's specific questions.

In a blog post about some of Sora's technical development, OpenAI said the model was partly trained on "publicly available data, mostly collected from industry-standard machine learning datasets and web crawls."

Brownlee's big questions for OpenAI

Brownlee threw dozens of prompts at Sora, asking it to generate videos of pretty much anything he could think of, including a tech reviewer talking about a smartphone while sitting at a desk in front of two displays.

Sora's rendering was believable, down to the reviewer's gestures. But Brownlee noticed something curious: Sora added a small fake plant in the video that eerily matched Brownlee's own fake plant.

The YouTuber showed all manner of "horrifying and inspiring" results from Sora, but this one seemed to stick with him. The plant looks generic, to be sure, but for Brownlee it's a reminder of the unknown behind these tools. The models don't create anything fundamentally novel; they're predicting frame after frame based on patterns they recognize from source material.

"Are my videos in that source material? Is this exact plant part of the source material? Is it just a coincidence?" Brownlee said. "I don't know." BI asked OpenAI about these specific questions, but the startup didn't address them.

Brownlee discussed Sora's guardrails at some length. One feature, for example, can make videos from images that people upload, but it's pretty picky about weeding out copyrighted content.

A few commenters on Brownlee's video said they found it ironic that Sora was careful to steer clear of intellectual property — except for that of the people whose work was used to produce it.

"Somehow their rights dont matter one bit," one commenter said, "but uploading a Mickeymouse? You crook!"

In an email to BI, Brownlee said he was looking forward to seeing the conversation evolve.

Millions of people. All at once.

Overall, the YouTuber gave Sora a mixed review.

Outside of its inspiring features — it could help creatives find fresh starting points — Brownlee said he feared that Sora was a lot for humanity to digest right now.

Brownlee said the model did a good job of refusing to depict dangerous acts or use images of people without their consent. And though it's easy to crop out, it adds a watermark to the content it makes.

Sora's relative weaknesses might provide another layer of protection from misuse. In Brownlee's testing, the system struggled with object permanence and physics. Objects would pass through each other or disappear. Things might seem too slow, then suddenly too fast. Until the tech improves, at least, this could help people spot the difference between, for example, real and fake security footage.

But Brownlee said the videos would only get better.

"The craziest part of all of this is the fact that this tool, Sora, is going to be available to the public," he said, adding, "To millions of people. All at once."

He added, "It's still an extremely powerful tool that directly moves us further into the era of not being able to believe anything you see online."

• McKinsey helps banks and financial institutions with their generative AI efforts.
• It outlined the dos and don'ts of seeing a return on AI investments in a report.
• Business Insider spoke with McKinsey's Larry Lerner about what will separate winners and losers.

The bill is coming due for Wall Street banks' AI investments.

It's been two years since generative AI captured the attention and dollars of bank leaders. They amassed teams of technologists to experiment with generative AI and run proofs of concepts. Some of those have since scaled to enterprise-wide initiatives used by thousands of employees. Now, leaders are beginning to question when these investments will pay off.

"That is the $20 billion question," according to Larry Lerner, a partner in McKinsey's banking practice.

For a handful of firms, Lerner said tangible returns are starting to emerge in the form of current cost savings, future cost avoidance, and incremental revenue. But for many, the reality is "POC purgatory," Lerner said, referring to proofs-of-concept pitfalls where firms get stuck in the experimentation phase and "become very tepid about really leaning in." In those cases, the "institution has spent the last two years investing and investing and not seeing anything at all," Lerner said.

According to an October report from Evident AI, which tracks AI adoption in financial services, only six out of 50 banks disclosed dollar-level cost savings or revenue lifts as a result of their AI investments.

So, what separates the frontrunners from the laggards? According to fresh research from McKinsey, it can come down to a few key decisions around concentrating efforts on a couple of uses, having CEO buy-in, and using generative AI in conjunction with other technologies. Most of all, it'll involve a mindset shift where AI is viewed and treated as a business opportunity rather than a technological problem, Lerner said.

Lerner outlined what will separate the winners from the losers. He declined to comment on specific companies.

Viewing AI as a business problem, not a tech one

Leadership teams have to recognize that generative AI is a business opportunity, not just a technology play, Lerner said. Because of that, he said business leaders should bear the brunt of the accountability, rather than that responsibility falling solely on tech leaders' shoulders.

"The institutions that make business leaders accountable for delivering their results will over time tend to do better because there's a much stronger partnership," Lerner said.

Concentrating firepower

Generative AI has lead to more value when there are only a handful of use cases, instead of every business unit doing a little bit here and there and seeing what sticks, Lerner said.

"Instead of having 60 use cases across 15 different business lines and functions, narrow down to three areas where you want to go deep," where you're reimagining the entire domain or workflow has led to a faster path to value, Lerner said.

Choose areas where ROI can actually be tracked

It's becoming increasingly clear that generative AI's main strength in saving workers time can't always be traced back to bottom-line impact, which is leading to some frustration in the boardroom.

"The value of what you're doing depends on how you're going to repurpose your time, and that's really hard to do," Lerner said. "Because it's an indirect sort of lever, it's very difficult to actually measure and get people to agree that there's value."

On the other hand, AI tools like call-center copilots and AI-powered marketing campaigns that improve the customer experience can generate incremental value that is measurable, Lerner said. One large bank referenced in the McKinsey report is projecting a 10% revenue increase thanks to a new analytics platform to target new customers and cross-sell products to existing ones.

For buy now, pay later fintech Klarna, leveraging an OpenAI-powered call center agent is estimated to bring some $40 million in profit this year, the company said in a blog post earlier this year. At the time, the AI was doing the work of 700 full-time agents, according to Klarna.

Lerner said he's starting to see some banks modify forward-looking hiring plans, especially in the contact center, thanks to the increase in self-service and faster resolution times. "That cost avoidance is absolutely measurable," he said.

Reusability is key

Build something once and redeploy it a hundred times, Lerner said. Doing so can accelerate development times and let companies scale faster because the tool has already gone through the required risk, security, and compliance approvals, he said.

Execution will come down to adoption

Getting workers and customers to adopt a new way of doing something or a new technology is one of the most important parts of the value equation. It's an old challenge that banks have had with previous technology cycles. When it comes to AI, "most companies have done a pretty bad job of getting adoption to the level that's going to yield the results that they want to yield," Lerner said.

• Google's search share slipped from June to November, new survey finds.
• ChatGPT gained market share over the period, potentially challenging Google's dominance.
• Still, generative AI features are benefiting Google, increasing user engagement.

ChatGPT is gaining on Google in the lucrative online search market, according to new data released this week.

In a recent survey of 1,000 people, OpenAI's chatbot was the top search provider for 5% of respondents, up from 1% in June, according to brokerage firm Evercore ISI.

Millennials drove the most adoption, the firm added in a research note sent to investors.

Google still dominates the search market, but its share slipped. According to the survey results, 78% of respondents said their first choice was Google, down from 80% in June.

It's a good business to be a gatekeeper

A few percentage points may not seem like much, but controlling how people access the world's online information is a big deal. It's what fuels Google's ads business, which produces the bulk of its revenue and huge profits. Microsoft Bing only has 4% of the search market, per the Evercore report, yet it generates billions of dollars in revenue each year.

ChatGPT's gains, however slight, are another sign that Google's status as the internet's gatekeeper may be under threat from generative AI. This new technology is changing how millions of people access digital information, sparking a rare debate about the sustainability of Google's search dominance.

OpenAI launched a full search feature for ChatGPT at the end of October. It's also got a deal with Apple this year that puts ChatGPT in a prominent position on many iPhones. Both moves are a direct challenge to Google. (Axel Springer, the owner of Business Insider, has a commercial relationship with OpenAI).

ChatGPT user satisfaction vs Google

When the Evercore analysts drilled down on the "usefulness" of Google's AI tools, ChatGPT, and Copilot, Microsoft's consumer AI helper, across 10 different scenarios, they found intriguing results.

There were a few situations where ChatGPT beat Google on satisfaction by a pretty wide margin: people learning specific skills or tasks, wanting help with writing and coding, and looking to be more productive at work.

It even had a 4% lead in a category that suggests Google shouldn't sleep too easy: people researching products and pricing online.

Google is benefiting from generative AI

Still, Google remains far ahead, and there were positive findings for the internet giant from Evercore's latest survey.

Earlier this year, Google released Gemini, a ChatGPT-like helper, and rolled out AI Overviews, a feature that uses generative AI to summarize many search results. In the Evercore survey, 71% of Google users said these tools were more effective than the previous search experience.

In another survey finding, among people using tools like ChatGPT and Gemini, 53% said they're searching more. That helps Google as well as OpenAI.

What's more, the tech giant's dominance hasn't dipped when it comes to commercial searches: people looking to buy stuff like iPhones and insurance. This suggests Google's market share slippage is probably more about queries for general information, meaning Google's revenue growth from search is probably safe for now.

So in terms of gobbling up more search revenue, ChatGPT has its work cut out.

Evercore analyst Mark Mahaney told BI that even a 1% share of the search market is worth roughly $2 billion a year in revenue. But that only works if you can make money from search queries as well as Google does.

"That's 1% share of commercial searches and assuming you can monetize as well as Google — and the latter is highly unlikely in the near or medium term," he said.

• Amazon has faced repeated delays in launching a new AI-powered Alexa.
• Integration with partners like Uber and Ticketmaster has complicated troubleshooting processes.
• Latency and compatibility issues have also caused delays.

Amazon's Alexa seems like the perfect product for the generative AI era.

Getting this powerful technology to actually work well with the digital assistant is a monumental challenge that's been plagued by gnarly technical problems and repeated delays.

Customer-friction concerns, partnership hiccups, compatibility questions, latency problems, and accuracy issues have snarled progress, according to internal Amazon documents and multiple people involved in the project.

The Alexa team is under immense pressure to get something out. A decade ago it launched with Echo speakers and became a household name. But that early success fizzled and the business has so far failed to become profitable, leading to drastic cutbacks and layoffs in recent years.

Some company insiders consider this AI moment to be a seismic opportunity for Alexa, and potentially the last chance to reignite consumer interest in the voice assistant through the power of large language models.

A product of this scale is "unprecedented, and takes time," an Amazon spokesperson told Business Insider. "It's not as simple as overlaying an LLM onto the Alexa service."

"RED" warning

One of the main challenges facing the new Alexa relates to how the digital assistant will interact with other companies and services, and who is responsible for customers if their requests, orders, and payments don't go smoothly.

In late August, Amazon was working on integrating 8 third-party applications, including Uber and Ticketmaster, into the upcoming AI-powered Alexa to handle various user inquiries.

At that time, the goal was to launch the new Alexa around mid-October, according to one of the internal documents obtained by Business Insider. However, it was still unclear which companies would be responsible for customer support issues, like payment and delivery errors, this document stated.

The lack of clarity could cause Amazon to send "frequent" customer contacts to the partner companies. Then, those partners would sometimes redirect the users back to Amazon, the document explained.

"This level of support would cause significant customer friction, when some of the orders/purchases are time-sensitive (meal orders or rideshare trips) and purchase mistakes can be expensive (e.g. buy Taylor Swift tickets)," the document said, assigning it a "RED" warning.

Release dates pushed back

Snafus like this have caused Amazon to push back the release date, almost on a weekly basis, according to some of the people involved in the project, which has been codenamed "Banyan" or "Remarkable Alexa." BI's sources asked not to be identified because they're not authorized to talk to the press.

For example, without more clearly defined responsibilities with third-party partners, Amazon expected further delays in the launch. "Alignment on customer support plans between Product teams and the 3P partners may push this timeline further out if any delays occur," one of the documents warned.

The company had once planned for a June launch, but after repeated delays, it told employees late last month that the new Alexa would launch "no earlier" than November 20, one of the documents said.

A few of people BI spoke with recently are even talking about the Alexa upgrade rolling out in early 2025, which would miss the key holiday period. Bloomberg earlier reported on a 2025 launch plan.

As of late October, Amazon had not settled on an official brand for the updated voice assistant, and instructed employees to simply call it the "new Alexa" until further notice, one of the documents said.

Alexa's huge potential

To be sure, Alexa has significant long-term potential in the generative AI era — as long as Amazon can iron out problems relatively quickly.

Time is of the essence, partly because the existing Alexa business has lost momentum in recent years. According to a recent report from eMarketer, user growth for major voice assistants, including Alexa, has declined significantly in recent years.

The sudden rise of ChatGPT has showcased what is possible when powerful AI models are integrated smoothly with popular products that consumers and companies find useful.

Some Amazon leaders are bullish about the AI-powered Alexa and a new paid subscription service that could come with it. At least one internal estimate projected a 20% conversion rate for the paid subscription, one of the people said. That would mean that out of every 100 existing Alexa users, roughy 20 would pay for the upgraded offering. Amazon doesn't publicly disclose the number of active Alexa users but has said it has sold more than 500 million Alexa-enabled devices.

An internal description of the new Alexa shows Amazon's grand ambitions: "A personalized digital assistant that can handle a wide range of tasks, including drafting and managing personal communications, managing calendars, making reservations, placing orders, shopping, scouting for deals and events, recommending media, managing smart home devices, and answering questions on virtually any topic," one of the documents said.

Customers will be able to access the new Alexa "through natural language using voice, text, email, shared photos, and more across all their devices like Echo, Fire TV, mobile phones, and web browsers," it added.

Amazon CEO Andy Jassy shared a similar vision during last month's earnings call, saying the new Alexa will be good at not just answering questions, but also "taking actions."

In an email to BI, Amazon's spokesperson said the company's vision for Alexa is to build the world's "best personal assistant."

"Generative AI offers a huge opportunity to make Alexa even better for our customers, and we are working hard to enable even more proactive and capable assistance on the over half a billion Alexa-enabled devices already in homes around the world. We are excited about what we're building and look forward to delivering it for our customers," the spokesperson said.

Smaller AI models

Still, the project has grappled with several challenges, beyond customer friction and partnership problems.

Latency has been a particularly tough problem for the AI Alexa service. In some tests, the new Alexa took about 40 seconds to respond to a simple user inquiry, according to three people familiar with the test results. In contrast, a Google Search query takes milliseconds to respond.

To speed up, Amazon considered using a smaller AI model, like Anthropic's Claude Haiku, to power the new Alexa, one of the people said. But that dropped the quality and accuracy of the answers, leaving Amazon in limbo, this person said. In general, smaller language models generate quicker responses than larger models but can be less accurate.

Amazon had initially hoped to use a homegrown AI model, one of the people said. Last year, Alexa head scientist Rohit Prasad left the team to create a new Artificial General Intelligence group at Amazon. The stated goal of the new team was to create Amazon's "most expansive" and "most ambitious" large language models.

However, this AGI team has not produced notable results so far, which led Amazon to consider Anthropic's main Claude offering as the primary AI model for the new Alexa, this person said. Reuters previously reported that Amazon was going to mainly power Alexa with Claude.

Amazon's spokesperson said Alexa uses Amazon Web Services's Bedrock, an AI tool that gives access to multiple language models.

"When it comes to machine learning models, we start with those built by Amazon, but we have used, and will continue to use, a variety of different models — including Titan and future Amazon models, as well as those from partners — to build the best experience for customers," the spokesperson said.

The spokesperson also added a note of caution by highlighting the difficulties of successfully integrating large language models with consumer applications. These models are great for conversational dialogue and content creation, but they can also be "non-deterministic and can hallucinate," the spokesperson added.

Getting these models "to reliably act on requests (rather than simply respond) means it has to be able to call real-world APIs reliably and at scale to meet customer expectations, not just in select instances," the spokesperson explained.

New risks

In late August, Amazon discovered several new risk factors for the AI Alexa service.

Only 308 of more than 100,000 existing Alexa "skills," or voice-controlled applications, were compatible with the new Alexa, presenting a "high risk for customers to be frustrated," one of the documents explained.

Some older Echo devices would not be able to support the AI-powered Alexa, the document also warned. And there were no plans for expanding the new service to dozens of overseas markets where Alexa is currently available, leaving a large user base out of touch, it also noted. Fortune previously reported some of these risk factors.

Integration headaches

As of late August, Amazon had 8 "confirmed" partner companies to handle certain tasks for the new Alexa, as BI previously reported. The company hopes to onboard roughly 200 partners by the third year of the new Alexa's launch, one of the documents said.

Integrating with some of these companies has already created headaches. One document said that Amazon struggled to develop a consistent troubleshooting process across every partner service. Companies including Uber, Ticketmaster, and OpenTable have deprecated their existing Alexa skills, further disconnecting them from the voice assistant.

Amazon's spokesperson said that, as with any product development process, a lot of ideas are discussed and debated, but "they don't necessarily reflect what the experience will be when we roll it out for our customers."

Amazon has also anticipated customer complaints, at least in the early launch phase. One internal document from late August stated that the new Alexa was projected to receive 176,000 customer contacts in the first three months of its release. At one point, Amazon considered launching a new automated troubleshooting service for issues related to its devices and digital services, including Alexa, according to one of the internal documents. That was later shelved.

Do you work at Amazon? Got a tip?

Contact the reporter, Eugene Kim, via the encrypted-messaging apps Signal or Telegram (+1-650-942-3061) or email ([email protected]). Reach out using a nonwork device. Check out Business Insider's source guide for other tips on sharing information securely.

• Modern factories, supply chains and Amazon have turned 'stuff' into a commodity. 
• The same inevitable supply-and-demand dynamic could wash over us again with generative AI.
• The ultimate outcome may be a new limited-edition luxury item: Humans. 

"Live experiences are the new luxury good," Kevin Hartz said in 2013 when Eventbrite, the ticketing startup he cofounded, got a big new funding round.

By that point, modern factories, supply chains, and Amazon had boiled down "stuff" to a commodity. You can now buy an overwhelming variety of tennis shoes, or spatulas, or sweatpants online. This abundance has taken much of the satisfaction away from purchasing physical things. This is why experiences, which by definition are finite, became more valuable.

There are only a few opportunities to see Taylor Swift on stage, versus the availability to purchase more than 20,000 kinds of tennis shoes on Amazon. So the price of Eras tickets soar, and shoes are cheap.

The same inevitable supply-and-demand dynamic is about to wash over us again with large language models and generative AI.

The ultimate outcome could be a new limited-edition luxury item: Humans. 

Unlimited content vs 'finite resources'

AI models can now automatically generate text, software code, medical diagnoses, images, voices, music, video, and lots more. The barriers to using this technology are falling away quickly. Anyone can fire up ChatGPT, GPT-4, DALL-E and other tools to produce an almost unlimited quantity of content.

This should be a boon to society. Many tasks will be completed more efficiently, making products and services more affordable and accessible, as venture capitalist Marc Andreessen recently explained.

There will be a reaction though: In a world of machine-generated abundance, human-centered services and experiences will become increasingly rare, valuable, and therefore desirable.

"The world's information is being turned into 1s and 0s and all this is being commoditized," Hartz told BI. "What can't be commoditized is finite resources like real estate, travel, seeing the sunset on Mediterranean, or surfing in Fiji. These are the luxury goods of the power elite."

Cooks, tutors, and robo-advisors

The more that AI automates restaurants, the more we'll want personal chefs such as John Barone, who cooks five days a week in the home of a wealthy Silicon Valley couple.

As AI tutor bots proliferate in education, the richest will pay for more exclusive access to the best human tutors for their kids.

The more robo-advisors handle our money, the stronger the urge of the wealthy to recruit savvy human experts to manage their family offices.

A new flood of automated emails

Email marketing is a simple example that some technologists are already worried about.

Generative AI tools are making it much quicker and easier to write marketing copy. The end result will be a flood of new emails that will overwhelm recipients and make them even less likely to open the messages.

"And our own machines will read those AI automated sales emails," Hartz quipped.

So, either your marketing email won't reach the humans you're trying to engage, or another AI bot will open it and you'll never be quite sure who read the message. A hand-typed email from a real human will be, relatively speaking, a rare and beautiful thing (complete with typos).

AI tutors versus small classrooms

AI models are beginning to revolutionize education, according to Sal Khan, the founder of Khan Academy. His organization has been working with OpenAI models to coach students in powerful new ways and help teachers develop class plans.

The gold standard throughout history has always been to have a personal tutor, and AI models can help personalize the education experience to bring some of this curated approach to more students, he explained during a No Priors podcast earlier this year.

"We don't have the resources to give everyone a tutor," he said during the podcast. "A generative AI tutor supporting students. That's going to be mainstream in 3 to 5 years," he added.

Pricey schools and a personal carpenter

And yet, Silicon Valley's top private schools, where many tech execs send their kids, are all about getting access to human teachers in small group settings. 

Castillja in Palo Alto highlights a student to faculty ratio of 7 to 1. Nueva, a Silicon Valley school for gifted kids, promises a similar ratio. The Menlo School in Menlo Park says it has a student-teacher ratio of 10 to 1 in the upper school.

These institutions cost $58,000 to $60,000 a year and I don't see any drop-off in demand among the tech elite. They're still jostling to get their kids into these bespoke, human-centered learning environments.

One persistent, apocryphal Silicon Valley story illustrates this point. On weekends, one tech billionaire has been known to hire a personal carpenter to hand-make wooden toys for their kids build and play with.

Who manages the money?

What about when it comes to managing fortunes amassed by successful tech entrepreneurs? The wealthiest rely on talented financial advisors who are hired directly to oversee this money in family offices.

Bill Gates has his own private investment firm, Cascade, which has been run by money manager Michael Larson since 1994. Elon Musk's family office, Excession, has been run by a former Morgan Stanley banker called Jared Birchall for years.

Using AI for trading has been tough so far. AI models are trained on masses of data from the past. When new situations arise, they struggle to adapt quickly enough.

Even quantitative hedge fund firms, which use machine learning and other automated techniques, rely on humans. Two Sigma, a famous quant firm, is for the first time exploring ways to add traders who rely on their human judgment to make money, Bloomberg reported recently.

"The major challenge with using things like reinforcement learning for trading is that it's a non-stationary environment," AI researcher Noam Brown said on the No Priors podcast in April. He's worked on algorithmic trading strategies in the past and was a researcher at Meta before recently joining OpenAI.

"So you can have all this historical data but it's not a stationary system," he explained, referring to how markets respond swiftly to world events and other developments.

Part of the problem relates to what he calls sample efficiency. Humans are good at learning quickly from a small amount of data, while AI models need mountains of information to train on.

"Humans are very good at adapting to novel situations," he added. "And you run into these novel situations pretty frequently in financial markets."

Social media bots vs. martial arts

AI is making social media increasingly machine-driven, too. Soon, human content creators will be vying for attention with content generated by AI models.

Last month, Meta CEO Mark Zuckerberg unveiled more than 25 new AI assistants with different personalities that use celebrities' images. Users will be able to interact with these bots on Meta's platforms in the future.

In a recent podcast, he described this new supply-and-demand situation well, saying human creators can't keep up with demand from followers.

"There are both people who out there who would benefit from being able to talk to an AI version of you," Zuckerberg explained. "You and other creators would benefit from being able to keep your community engaged."

So Meta will make an AI version of celebrities that can post constantly. Again, this will be infinite. And actually interacting with the real human celebrity will become more rare and valuable.

Meanwhile, when Zuckerberg is relaxing outside of work, he spends some of that time pursuing a very human pastime: Rolling around with other humans in martial arts contests.

Medical models and human doctors

AI models, such as Google DeepMind's Med-PaLM 2, are becoming incredibly good at answering medical questions and analyzing x-rays and other health data. But when wealthy parents have really sick children, they will still seek out the smartest doctors in the relevant fields of medicine.

You can see this in Silicon Valley's embrace of medical concierge services that provide special access to doctors and other human health specialists.

One Medical succeeded by offering better access to human doctors, and Amazon ended up buying it for almost $4 billion.

"We're inspired by their human-centered, technology-forward approach," an Amazon executive said when the deal was announced.

'Utility, value and signaling'

Hartz, a venture capitalist who now chairs Eventbrite's board, says successful technologists will continue to spend heavily on human experiences. But he says this depends on the activity and the motivations behind different actions.

He breaks this into "utility, value and signaling."

Many standard, common situations can be handled by software bots or even physical machines. Repetitive tasks at work and some educational functions are examples of these utility-type solutions.

In other situations, users will get more value from having machines handle the work, so humans can focus on more valuable tasks. If you're a well-paid machine-learning engineer, it will be better to have a robot clean your house so you can focus more on your job, he explained.

And then there will still many situations where humans will want to enjoy their success and signal the fruits of their achievements. And these activities will increasingly focus on finite human resources and experiences, Hartz said.

"You can't put on headset and pretend to be in Fiji," he added.

• Business Insider asked 27 venture-capital investors to nominate the most promising fintechs.
• Fintechs using AI to help Wall Street firms, bankers, and consumers lead this year's series.
• Here are the top 15 AI fintechs, according to VCs.

Fintech investors still see at least one bright spot in the industry, despite funding to the sector hitting one of its lowest points since the pandemic.

Total funding to fintechs fell once again last quarter, according to CB Insights' third-quarter data. Only 753 deals were inked, notching the lowest quarterly level since 2017.

However, the dealmaking drought could ease up in the next year as a result of antitrust scrutiny softening and VCs might be more willing to open up their pocketbooks. One area that investors will likely hone in on once they do is AI.

Earlier this year, Business Insider asked dozens of VCs to identify the most promising fintechs to watch. Nearly one-quarter of the startups they named are leveraging AI as a key part of their offering. Indeed, it is difficult to point to one area of finance where AI startups aren't threatening to change the way people bank, invest, save, and work.

Some of the startups on this list are business-facing, helping dealmakers negotiate debt agreements, streamlining tedious processes for junior bankers, or automating manual processes for accountants and CFOs. Others use AI to serve consumers, whether it's helping them figure out the best way to pay off debts or maintaining access to healthcare between jobs.

The startups named haven't raised beyond a Series C and include a mix of investors' portfolio companies and ones they have no financial interest in.

Here are 15 of the most promising AI fintechs to watch, according to top VCs.

BeatBread

Cited by: Deciens Capital (investor)

Total raised: More than $150 million

What it does: BeatBread uses AI to analyze and predict revenue potential for the music industry, providing funding advances to a broad range of artists.

Why it's on the list: "Artists of all sizes want independence and ownership over their music, to work with their preferred partners, and to control their own destinies. Historically, there hasn't been a real alternative to the major label advance for artists to get the capital they needed to scale their careers, which locks them into the label ecosystem," Dan Kimerling, the managing partner at Deciens Capital, said.

"2024 has been a pivotal year for BeatBread, marked by strategic moves and partnerships that further solidify its mission," Kimerling said, referring to its partnerships with the administrative publishing company Kobalt and its subsidiary AMRA to offer artists increased royalties and faster payments. Other strategic moves include a series of deals providing funding to independent labels to expand how BeatBread provides capital to artists.

Brico

Cited by: TTV Capital, Homebrew

Total funding: $8.1 million

What it does: Brico helps financial institutions and fintechs manage their licensing by using AI to automate applications and renewals.

Why it's on the list: "With Brico, businesses can effortlessly navigate the complexities of acquiring, renewing, and managing compliance for various financial licenses — including Credit, Money Transmitter, Mortgage Loan Originator, and more — in all 50 states," Lizzie Guynn, a partner at TTV Capital, said. "Brico makes regulatory compliance seamless and cost-effective with its user-friendly tools that reduce time and money spent on financial licenses."

"It's addressing a very manual and expensive process that nearly every financial services company needs to deal with on an annual basis," Satya Patel, a partner at Homebrew, said.

Cascading AI

Cited by: QED Investors, Vesey Ventures

Total raised: $4.1 million

What it does: Cascading AI, through its main product Casca, offers loan-origination software for the banking sector with an integrated AI assistant that allows firms to extend their hours.

Why it's on the list: "Customers do not operate on the 9-to-5, Monday-to-Friday schedules that banks do," Laura Bock, a partner at QED Investors, said. "When a pizzeria's oven breaks, the owner is inquiring about a loan after closing shop. While today, it might take nearly three days to hear back from a loan officer after submitting an application, financial institutions using Casca's AI platform are able to unlock 24/7, 365 support for current and potential customers."

Dana Eli-Lorch, a founding partner at Vesey Ventures, said: "Their flagship product, an AI-powered loan assistant, enables manifold increases to banks' productivity and loan conversion rates, all while enhancing both accuracy and applicant experience. Casca exemplifies the powerful impact AI can have on financial services, driving significant operational efficiency and customer satisfaction."

Clerkie

Cited by: Flourish Ventures (investor)

Total raised: $41 million

What it does: Clerkie embeds its AI debt-automation software in financial institutions' mobile apps, allowing consumers to make financial decisions about their debts and discover solutions if they're struggling to pay them off.

Why it's on the list: "Clerkie's data flywheel and network create a win-win scenario for both consumers and financial institutions. Consumers enjoy a seamless experience within their banking app, with flexible solutions tailored to their specific cash flow needs, helping them avoid the collections process and protect their credit scores. Banks benefit from direct ROI through loan repayment while maintaining customer relationships," while also expanding loan-to-value ratios, Flourish Ventures' Emmalyn Shaw said.

She added that "Clerkie assumes no balance-sheet risk, serving as the debt-network and debt-payment infrastructure for financial institutions."

Comulate

Cited by: Pathlight Ventures (formerly Exponent Founders Capital)

Total raised: About $5 million

What it does: Comulate automates insurance statement processing, reconciliation, revenue recovery, and forecasting.

Why it's on the list: "Leveraging AI to drive real revenue lift for insurance carriers is driving success in a category" that's historically been hard to break into, Charley Ma, a cofounder and managing partner of Pathlight Ventures, said.

Coris

Cited by: Pathlight Ventures (investor)

Total raised: $3.7 million

What it does: Coris builds software for fintechs and other tech companies to manage risk and fraud among small- and medium-size business clients.

Why it's on the list: "Aggregating unstructured data on SMBs to generate insights at scale is challenging. Coris is at the forefront, leveraging a variety of methods across LLMs, ML, and good old-fashioned software to establish itself as the leading platform for managing SMB risk and fraud — already working with clients like Mindbody and ClassPass," Pathlight's Ma said.

Fintary

Cited by: Harlem Capital (investor)

Total raised: $2.5 million

What it does: Fintary helps insurance companies manage their finance and accounting needs by using AI to automate workflows.

Why it's on the list: "They have been invited to their customers' conferences in order to share the product with their customers' customers," Henri Pierre-Jacques, the cofounder and managing partner of Harlem Capital, said. He said Fintary has grown more than 10 times since Harlem's investment last fall, adding that "the quick ramp has been one of the fastest we've seen for a preseed company."

Greenlite

Cited by: Greylock (investor)

Total raised: $4.8 million

What it does: Greenlite automates compliance processes using AI for fintechs and banks.

Why it's on the list: "Greenlite has seen exceptional customer demands with enterprise banks and fintechs and has proven one of the few enterprise-grade applications for generative AI — automating tedious compliance workflows like alert handling, periodic reviews, and document processing, improving efficiency and reducing human error," Seth Rosenberg, a general partner at Greylock, said.

Iris Finance

Cited by: Redpoint Ventures

Total raised: $3.5 million

What it does: Iris Finance offers consumer-facing companies AI-powered financial planning and analysis software.

Why it's on the list: "While the notion of AI bookkeeping is very much in vogue today, replacing Quickbooks is hard — and not something most brands or outsourced accountants are looking to do in the near term," Redpoint's Clark said. "Iris, instead, complements Quickbooks with a more holistic AI-powered CFO-in-a-box for brands, enabling them to seamlessly track and improve day-to-day sales and margin performance across channels, which much more closely aligns with what founders want and how modern brands are managed."

Materia AI

Cited by: Bain Capital Ventures

Total raised: $6.3 million

What it does: Materia AI helps accountants organize their data, enabling them to automate parts of their work.

Why it's on the list: "With a decline in new auditors and an immense volume of manual data entry, professional-service audits are the perfect place for an AI copilot," Alysaa Co, principal at Bain Capital Ventures, said. "LLMs enable the automation of work like ingesting large sets of unstructured financial data, searchability, comparing against historicals and across the industry, and direct citations for where the data comes from."

Nilus

Cited by: Vesey Ventures (investor)

Total raised: $8.6 million

What it does: Nilus offers an AI-powered cash and treasury management platform for fintechs, financial firms, marketplaces, and other companies moving money.

Why it's on the list: Nilus "provides better data connectivity combined with AI to transform the CFO suite: a trend we are actively investing behind," Lindsay Fitzgerald, a general partner and the cofounder of Vesey Ventures, said.

"With Nilus, treasurers can skip the manual reconciliation work that previously took most of their day and focus on actions that can drive bottom-line impact. We think Nilus is poised to become the default software for modern treasury teams, displacing decades-old workflow tools like Kyriba and GTreasury," she said.

Noetica

Cited by: Avid Ventures, Index Ventures

Total raised: $7.85 million

What it does: Noetica helps deal professionals negotiate debit agreements with their data using an AI platform that benchmarks terms in corporate debt transactions.

Why it's on the list: "Noetica is a capital-markets data company for corporate debt, a market valued at trillions of dollars. Its AI-powered software allows professionals to upload any credit or bond document and compare all terms to similar public and private deals," Jahanvi Sardana, a partner at Index Ventures, said.

"Corporate debt terms are time-consuming and difficult to benchmark, leading deal professionals, such as lawyers and investment managers, to often miss higher-risk terms, as well as opportunities for negotiation. By building the largest proprietary dataset of corporate debt terms, Noetica is changing how these deals are negotiated and transacted," Tali Miller, a founding investor at Avid Ventures, said.

Novella

Cited by: Avid Ventures (investor)

Total raised: $2.5 million

What it does: Novella is an AI-powered insurance wholesaler specializing in excess and surplus insurance, which addresses higher-risk situations that standard carriers don't usually cover.

Why it's on the list: "Given the complexity of E&S insurance, it is sold through wholesalers who have relationships with specialty carriers, and retail brokers must work with these wholesalers to access these carriers. However, brokers have been frustrated by the leading wholesalers such as Ryan Specialty, Amwins, and CRC Group, whose lack of technology and system integrations lead to slow, inefficient, and opaque quoting processes," Avid Ventures' Miller said.

"The E&S market continues to grow," Miller said, adding that E&S direct premiums written in the US climbed to more than $86 billion in 2023, more than doubling since 2018.

"Using data and AI, Novella aims to reinvent this massive industry by making the information transfer between brokers and carriers fast and error-free and, ultimately, automating quote creation," she added.

Rogo

Cited by: Two Sigma Ventures

Total raised: $26 million

What it does: Rogo is building a generative AI assistant to help investment bankers and analysts do their jobs more efficiently.

Why it's on the list: "Rogo's platform is purpose-built for the complex data needs of the financial sector, allowing nontechnical users to query vast amounts of financial data using natural language processing. This is a game changer for institutions like banks, investment firms, and insurers," Frances Schwiep, a partner at Two Sigma Ventures, said.

"I see immense potential in Rogo's ability to give first-of-its-kind access to critical financial analytics, positioning them as a key player in transforming how financial institutions interact with their data to drive more informed decisions across the industry," she added.

See the pitch deck for Rogo's $7 million seed.

When

Cited by: TTV Capital (investor)

Total raised: $7 million

What it does: When uses an AI assistant to help exiting employees maintain access to healthcare by providing affordable alternatives to COBRA and making it easy to compare pricing and deductibles.

Why it's on the list: "There are more than 700,000 companies in the United States with 20-plus employees, which means they are required by law to offer COBRA. Last year's 721,677 planned job cuts brought some of the largest reductions in company head count that we've seen in the past two decades," TTV Capital's Guynn said.

"Offering an alternative to expensive, inflexible COBRA not only makes common sense but also economic sense. COBRA participants are three times more costly than active employees, which is especially burdensome for self-insured companies. To date, companies that offer When's fixed-dollar health-insurance premium reimbursement have seen an 80% conversion rate from COBRA. Employees that applied their When benefit to available plans have saved as much as 50% in out-of-pocket healthcare costs," Guynn said.

• Finance firms are keen on AI agents that can automate combinations of tasks.
• Demand for AI agents is giving birth to a new class of startups and VCs hungry to invest in them.
• It was a topic of conversation at the Evident AI Symposium in New York on Thursday.

"Talk to this like a teammate and treat it like a teammate."

That's Danny Goldman's guidance to private-equity customers of his startup, Mako AI, which offers a generative AI assistant for junior finance professionals and is backed by Khosla Ventures, an early investor in OpenAI.

His hope is that "engaging with Mako looks much more like engaging with a real human associate than a software tool," he previously told BI. Goldman, who worked in private equity before cofounding Mako AI, predicts that in a year or two, every junior on Wall Street will have their own AI direct report.

It's not just juniors, either. JPMorgan CEO Jamie Dimon, is a "tremendous user" of the bank's generative AI assistant suite. Teresa Heitsenrether, JPMorgan's chief data and analytics officer, said at a conference last week that JPMorgan is working toward giving employees AI assistants that are specific to them and their jobs.

Wall Streeters, say hello to your new coworker. Across the industry, AI agents are beginning to permeate the labor force as assistants who can help humans prep for meetings, write their emails, and wade through troves of information to answer questions almost instantaneously.

In many cases, AI agents are still limited to specific, individual tasks like querying internal data and creating PowerPoints and emails. To take AI agents a step further, technologists and startup investors are fueling a shift to so-called multi-agent systems that coordinate several AI agents to complete more complex tasks more autonomously.

Some tech executives at the Evident AI Symposium said they could see a world with more artificial intelligence agents than humans by 2025. But what will work and life look like in an increasingly hybrid world with humans and bots? Well, that's still being worked out, according to a number of tech executives at the Evident AI Symposium Thursday.

"What's really exciting about agents is that we are still figuring out the tasks they're actually good at, the tools they know how to use, the tools we have to teach them how to use," said Gabriel Stengel, cofounder and CEO of Rogo, which is building the generative AI equivalent of a junior banker.

Another question that still needs to be answered is how to define when an agent is smarter or not than a human, said Kristin Milchanowski, chief AI and data officer of BMO Financial Group.

To some extent, benchmarking humans against AI agents is already happening. In a recent University of Cambridge study that compared who could run a business better, AI outperformed humans on most metrics including profitability, product design, and managing inventory. But they fell short when it came to making decisions on the fly.

Heitsenrether, speaking at the Evident AI conference, told the audience that, over time, she expects AI to be seamlessly embedded in an employee's workflow. By this time next year, she said that she hopes to have a clearer picture of what a more personalized AI assistant for each employee might look like.

But unlocking more autonomous uses of AI is going to require more than technological breakthroughs.

"We don't have a lot of trust right now in these systems," Sumitra Ganesh, a member of JPMorgan's AI research team, said at the symposium.

"We have to slow-walk it to release it to people who are experts who can verify the output and go, 'Okay, that looks fine, you can take that action,'" Ganesh said. "But that's kind of babysitting these agents at this point," she added. "But hopefully, it's like training wheels — at some point, we will be confident enough to let them go."

• Jeff McMillan is Morgan Stanley's head of firmwide AI.
• His team vets all new AI pitches, but business leaders ultimately decide what to use.
• He outlines the multiple steps involved from pitching an idea to getting it into production.

If Jeff McMillan does his job right, it will look very different in three years.

"Think about it. We don't have a head of PowerPoint at Morgan Stanley or Excel," McMillan told Business Insider. "These are just enabling technologies," he added.

He was named Morgan Stanley's head of firmwide AI in March to help integrate the technology across the firm. While much of his job these days is focused on getting businesses up to speed with AI and implementing it efficiently across the bank, he said his ultimate goal is for the technology to be ingrained into workers' everyday lives.

Since his promotion, McMillan has led the rollout of a few generative AI tools in the bank's wealth-management division, and has more use cases in the pipeline, he said. The bank's push into generative AI has been fueled by its early partnership with ChatGPT-maker, OpenAI, and coincides with Wall Street's recent obsession with generative AI to boost productivity and reduce grunt work for workers.

McMillan encourages employees to pitch new AI solutions. His firmwide team acts as a filter and vets the ideas, which can come from practically anyone who's done the required training at the bank. To avoid creating an unruly situation where thousands of technologists, analysts, and bankers are building their own AI tools, he's devised a rigorous multi-step process that involves pitching solutions to some of the firm's top execs and devising a business value proposition.

As part of his role, McMillan co-chairs an AI steering group formalized earlier this year, with Global Director of Research Katy Huberty. The steering committee, which has representatives from each department, vets all AI use cases pitched by employees.

The steering group is working through more than 30 use cases that are in various stages on the way to launch, McMillan said. AskResearch, an assistant that gives investment bankers, salespeople, and traders information found buried in tens of thousands of research reports, is the latest generative AI product to make it through the process since launching McMillan's team.

Many of the pitches the steering committee sees fall into two buckets: use cases that are relevant to several groups, or that matter to a specific team or group of users. For the former, McMillan is able to coordinate teams across the firm to collaborate and build solutions together, with the aim of increasing reusability.

By structuring the AI approval process this way, McMillan hopes to enable the bank to innovate without sacrificing safety.

"While there might be creative tension between experimentation and process, I believe that a rigorous process will ultimately allow us to develop and deploy technology faster and more efficiently," McMillan said.

Inside the 8-step process

Although pitching AI solutions is open to anyone at the firm, there is some leg work involved. Mainly, workers have to complete specific training on governance and AI principles and meet standards around information security.

The AI steering group meets every other week to listen to the pitches, usually going through five or six proposals. The steering group usually either approves or approves with conditions, like rethinking an aspect of the solution or coming together with other teams that pitched similar ideas. In some cases, pitches are rejected — something McMillan says he generally tries not to do.

"I don't want to be in a position where I'm telling people no. I want to tell people yes, and this is the best way to get to it," McMillan said.

For presentations that are approved, the next steps typically involve identifying the people who will execute and figuring out who from tech, legal, compliance, and risk needs to be involved. Workers going through this process also have to articulate deliverables and identify the risks, as well as having a plan for mitigating those risks. That might be a standard set of questions and answers used in testing or making certain teams aware of the potential risk.

They will also have to put together a business value proposition that outlines the benefits, which could be quantifiable, such as decreasing margin or operating costs, finding new revenue streams, or decreasing risk.

Every other week, the AI steering committee meets to review the status of these projects. At the end of the process, the group pitching presents a final time to the steering group for go-live approval to ensure all the conditions are met. Finally, use cases go into production.

"What we're doing is we're helping them prioritize. We're grouping them, and then my team, we handhold you. We say, okay, what are you trying to do? We help you set up the environment. We make sure you've got the right level of APIs, we are by your side as you work with our legal, compliance, and risk process," McMillan said of his business partners.