The Best Place to Plug In AI: Corporate Finance

01 /The Concept

Daloopa built an AI engine that constructs financial models from company financial data.

The basic mechanic: feed it a company's financials, and it builds a model of how that business actually works – one you can use to assess the durability of the business model and run forward projections.

The engine currently covers 3,000 public companies. Its source material is SEC filings, earnings call transcripts, and other public disclosures.

The platform's target audience is investment funds, banks, private equity firms, and sell-side research shops.

The interface is an Excel plugin. Daloopa builds the financial model for a given company directly into the analyst's spreadsheet, where they can work with it in the environment they already know. Every source figure in the model is traceable back to the original document it came from. The company's tagline – "trust, but verify" – doubles as a product philosophy: the AI does the heavy lifting, but every number has a source link so analysts can check the work.

When a company releases a new quarterly report, the analyst clicks one button to refresh the model. Custom calculations the analyst has layered on top update automatically – only the source data changes. If a new report contains a metric the company has never disclosed before, the plugin flags it so the analyst can factor it into their own work.

There's a free tier: analysts can explore Daloopa's models for five companies of their choice. The full platform is paid – pricing requires a direct conversation with the team.

A premium tier called Daloopa Plus delivers new financial data in real time as it appears on company investor pages or in SEC filings.

The startup is clearly targeting large institutional investors. The website prominently notes that nine of the 20 largest Disney shareholders use Daloopa, six of the 20 largest Uber shareholders use it, seven of the 20 largest Amazon shareholders use it – and so on.

Daloopa just raised $18M in new funding, bringing total investment to $41.4M.

02 /Why It Matters

First, there's an unusually honest line in the founders' funding announcement: "Daloopa is a simple business." We take public data, clean it, standardize it, put it in a database, and keep it updated. That framing took some nerve – most founders can't resist talking up the complexity of what they've built and the sophistication of the technology behind it.

The customer testimonials they chose are equally plain. "I used to spend three days building a financial model by hand, entering data and formulas. Now it's instantaneous." "After a new earnings release, I no longer have to sit and manually update everything. I click a button, refresh the data, and I'm home in time for lunch with my kids."

This reinforces a point worth making repeatedly: the most productive application of AI right now isn't solving exotic problems – it's eliminating the manual, repetitive work that people already do every day. Work that's slow, error-prone, and nobody particularly enjoys. You don't need a clever idea for an AI platform; you just need to find something that lots of people are still doing by hand and automate it.

But there's an even more interesting trend underneath all this, which is the coming era of AI-to-AI communication – systems that talk to each other directly, without humans converting data between formats in the middle.

Consider what's happening already. On one side, AI can expand a few bullet points into a full article. On the other, AI can compress a long article back down to bullet points. But if the same ideas get expanded and then re-compressed, something is always lost in translation. Why create that intermediate step at all?

Search and commerce show the same dynamic: AI tools already help sellers optimize listings, and AI tools help buyers parse search results. The logical endpoint is buyers' AI talking directly to sellers' AI – a point [covered previously](/review/bez-jetogo-oni-nichego-poleznogo-dlja-tebja-ne-sdelajut).

In financial analysis, the same pattern applies. AI (Daloopa) expands raw financial disclosures into structured models. Another AI reads those models to make investment decisions. Putting a human analyst in the middle of that data flow to manually maintain the model – which is what Daloopa replaces – looks, in retrospect, like an obvious bottleneck.

03 /Opportunities

Daloopa's model is clean and extensible: find a domain where experts spend significant time manually extracting and structuring data from documents, automate that extraction, and deliver the structured output directly into the tool experts already use.

Financial analysis is the beachhead, but the pattern works elsewhere. Legal contract analysis, regulatory compliance tracking, medical literature review, supply chain document processing – any field where professionals spend hours converting unstructured documents into structured models is a candidate. The Scaleup Finance model points to an adjacent direction: instead of just delivering models, build the layer that interprets them – moving from data structuring to insight generation.

The furthest extension of this logic is a marketplace connecting document-processing AIs (like Daloopa) with decision-making AIs (like investment algorithms), where the human's job is to define the strategy rather than manage the data. That's not today's product – but it's the trajectory Daloopa is on, and it's worth building toward.