Hello and welcome to Eye on AI. In today’s edition…an exclusive on a new startup vying to meet the financial sector's thirst for AI; Google drops its policy against creating AI that causes harm; DeepSeek ranks last in cybersecurity; Amazon gears up to finally release its AI-powered Alexa; and BigBear.ai surges after securing a defense deal.
Before every high stakes decision is made on Wall Street, analysts must scour vast information sources, compile data, and create often dense reports. This won’t be the case much longer if Model ML, a new company emerging today from stealth, has its way.
The company—also today announcing a $12 million funding round led by Y Combinator and LocalGlobe—is offering an AI-powered workspace that automates financial research and due diligence. Basically, it retrieves information from every data source a human at a firm has access to and creates whatever is requested, whether that be a company profile or 100-page memo.
“The general idea is, you never have to go and gather information,” cofounder and CEO Chaz Englander told me in an exclusive interview and demo.
The company comes as financial institutions are obsessing over how to tap AI to be more efficient, and as always, make loads more money. Startups like Kavout, Finomial, Alphasense, and more are all in the business of helping investors invest faster and better, and some like Samaya are similarly combining financial analysis and agents. Not to mention, the powerhouses of the AI industry are all moving into models for in-depth analysis and research, including OpenAI's new Deep Research tool, Google's Deep Research, and Anthropic's Clio.
How it works
Model ML was founded by brothers Chaz and Arnie Englander, serial entrepreneurs behind last-mile delivery company Fancy (acquired by Gopuff) and marketplace company Llama (acquired by Hygglo). Initially, they began working on the product for their own use to guide their own investments. People in their circles showed interest, so they started spinning it into a company in 2023.
When you open Model ML, it looks a lot like Google Drive. It has its copycat versions of Excel, Powerpoint, Word, etc., which ensures that no information ever needs to leave the workspace. Everything is hosted on a company or team’s own infrastructure, aiding in security and compliance. Of course, users can export to the real Excel, Word, and other file formats when they need to.
Model ML is also connected to every piece of software and information store a user has from emails and files to customer relationship management software (CRM) to any databases the user has access to. Users can input natural language prompts to have the system retrieve whatever information they need and present it however they’d like. For example, you can prompt it to create comparable company analysis charts, often referred to as comps. It will instantly start filling in the columns of information, linking to the original source for every bit of data (and highlighting the information in the original source so it can be easily cross-checked).
“It's effectively a tabular view on top of an agentic system,” Chaz Englander described, adding that popular use cases among early customers include creating earnings summaries, trend analysis, market mapping, querying structured info from huge document repositories, and more.
The other key, according to Englander, is how the use of templates makes it possible to streamline some of the most repetitive tasks in the industry. Analysts create comps day after day. Model ML not only fetches the info and creates the chart, but you only need to set it up once.
Job cuts loom on Wall Street
The appetite for efficiency in the financial sector is big and growing as firms flock to the new generation of AI tools. Englander says the company already has around 40 customers, including many of the largest financial organizations in the world. He said he couldn’t name any at this time, noting that these conversations are happening at the executive level and are at the forefront of firms’ larger strategies.
“It's not like they're choosing the next data vendor or the next CRM,” Englander said. “This is going to play such a fundamental part in their overall success as a business that people—and rightfully so—are mindful about how they're articulating themselves, what they're automating, and what they're not automating.”
Last month, a new report from Bloomberg Intelligence suggested that as many as 200,000 jobs on Wall Street could soon be lost, citing firms’ estimations around cutbacks in the coming years. On average, the chief information officers surveyed said they expect a 3% reduction in their workplaces over the next three to five years, with one in four predicting cuts of between 5% to 10%.
Of course, it’s the routine, repetitive tasks that are on the chopping block. It’s not a done deal yet, however: AI has historically struggled to accurately analyze data and spreadsheets, as my Eye on AI colleague Jeremy Kahn reported on earlier this week in an exclusive about a potential breakthrough in this area. But there’s a whole lot of effort going into, as I like to say, using AI to “make spreadsheets spreadsheet better,” and a lot of quick progress, too.
And with that, here’s more AI news.
Sage Lazzaro
sage.lazzaro@consultant.fortune.com
sagelazzaro.com
