Built inside an investment firm, for its own analysts.
Baseline scores more than 3,600 companies on the qualitative questions a quantitative screen cannot reach, and shows its reasoning on every one of them.
The problem we had
There are many quantitative tools in the market, but none effectively analyse the other qualitative side at scale. This work is done by analysts, one company at a time, often taking days. Baseline scores more than 3,600 companies on 30+ fields such as pricing power, competitive intensity and accounting quality, each with an analyst depth write-up behind, accelerating the due diligence process.
Why a fund built it rather than a software company
A5 Labs is separate to, but born out of A5 Capital, a London-based private investment firm. Baseline began as an internal tool, and the first users were the analysts sitting on the fund.
Every field on the platform exists because somebody needed to answer that specific question about a real position and could not do it at scale. That origin still sets our priorities. We are our own first users, and we build product that satisfies institutional investor workflow.
How it works
The same fixed set of questions is asked of every company. Pricing power. Barriers to switching. Mission criticality. Competitive intensity. Management quality. Accounting risk. More than thirty fields in total, each one scored from 1 to 5, where 5 is the most favourable reading.
Rationale
Contracted price escalators on a mission-critical product let the company reprice ahead of inflation, with no measurable churn across the last three renewal cycles.
Every field carries a written rationale underneath the number. That matters more than the score itself, because it means you can read the argument, check the evidence and disagree with it. A number on its own tells you what we concluded. The rationale tells you why, and lets you decide whether we were right.
How we check ourselves
Every score is produced by a language model, so the only honest question a professional investor can ask is how often it is wrong. We would rather answer that with a measurement than a reassurance.
- July 2026First blind test. Twelve companies drawn across marquee, mid-cap and thinly covered names, with claims verified against primary sources.
- ContinuousUniverse-wide integrity checks. Automated tests that read every company on the platform and flag any result whose text disagrees with its own record.
The team
Baseline is built by a small team inside A5 Labs, working with engineering partners in London and India.
Arnav Agarwal
Ex-Rothschild private equity
Harry Read
Ex-Rothschild/GS equity research
Working with us
Enterprise
Firms can run Baseline against their own prompts and their own methodology, so the questions asked of every company are the ones your process actually cares about. Results are available through the platform or directly inside your team's AI tools.
Everyone else
We are always glad to hear from people pushing at the same frontier. Email harry@baseline.a5labs.com with ideas, criticism, or a company you think we have scored wrongly.
Baseline surfaces and screens candidates. It does not predict returns.
A5 Labs, the incubation arm of A5 Capital.