Itoflow raises $2.5m for AI agents that run an investor’s own process

On 25 August, Itoflow, a London company building AI agents for professional investment teams, announced a $2.5m (£1.8m) pre-seed round led by Balderton Capital. Angel investors including Barney Hussey-Yeo, founder and chief executive of Cleo, joined the round, according to Balderton’s announcement. Itoflow will spend the money on its engineering and quantitative research teams, on product development and on what it calls its commercial and regulatory work.

The company is not yet three months old. Companies House records Itoflow Ltd as incorporated on 29 May 2026, with co-founder and chief executive Aditya Jha as its first director. Jha and Abinash Meher each hold between 25% and 50% of the shares. The investor was in place well before the news went out: Sivesh Sukumar, a principal at Balderton, was appointed a director on 30 June, and the company recorded an allotment of new shares on 10 July.

Balderton says Itoflow was founded by three graduates of the Indian Institute of Technology. Jha spent nine years as a quantitative researcher at JPMorgan, RBC and Tower Research Capital, where he was business head for mid-frequency trading, according to the announcement. The wider founding team built large-scale infrastructure at Apple, Google and Microsoft, and also brings quantitative research experience from JPMorgan.

Balderton set out its case for the deal in the announcement. “The most sophisticated investment firms spend years building the systems and research teams needed to test ideas, manage risk and respond as markets change,” Sukumar said. “Aditya and the team are changing the economics of that model, giving many more investment firms access to systematic capabilities that were previously too expensive and complex to build.” Itoflow is betting that the scarce resource in systematic investing is the research team rather than the strategy, and that agents can stand in for the team while the strategy stays with the client.

A quant desk without the headcount

The company says its agents learn how each firm researches opportunities, manages risk and reviews portfolios, without imposing an investment philosophy of their own. Each team’s approach becomes a set of governed workflows that can run continuously across equities, bonds, ETFs, commodities and digital assets.

In the current pilots, investment teams write down in plain English the approach, risk constraints and review criteria they want applied. The agents then run quantitative research and backtesting, monitor portfolios and flag material changes with supporting evidence. Itoflow says this lets teams review more investments and monitor more portfolios without building a much larger quantitative research function.

Control as the selling point

Much of what Itoflow describes is about limits on the agents rather than their reach. Customers set which actions and thresholds need human approval. Institutional customers can run the platform inside their own infrastructure, so sensitive data and proprietary strategies stay in their own environment. The company also built its agent infrastructure from scratch, which it says gives it more control over the security and reliability of agents meant to run for months or years as markets change.

For a regulated investment firm, an agent acting inside a client mandate without an audit trail or a human sign-off would be hard to get past a compliance team. The announcement says a good deal about approval and evidence, and nothing about investment performance.

Hussey-Yeo said much the same in the announcement: “Building Cleo has taught me that financial AI only works when it’s intuitive, trustworthy and keeps people in control.” He added that Itoflow “gives investment teams sophisticated tools without asking them to give up their own process or judgment.”

Pilots, but no named clients

Itoflow is running pilots with what it describes as a number of hedge funds and family offices globally, and is in discussions with exchanges, brokerages and financial-data providers. None of the pilot firms is named, and the company has not published revenue.

Jha’s stated ambition goes further than running a client’s rules. “Our long-term research goal is to build self-improving investment agents,” he said in the announcement. An agent supporting a mandate over months and years, he added, “should learn which approaches continue to hold up, identify promising new signals” and spot when an apparent edge has decayed.

An agent that finds new signals and drops decayed ones is changing the strategy, which is the part Itoflow says it leaves to the client. Jha frames it as discovery “within the parameters set by each investment team”, and the approval thresholds customers set are the mechanism the company offers for drawing it.

For a company incorporated in May, Itoflow has a good deal in place: a lead investor on its board within five weeks, a founder who headed mid-frequency trading at Tower Research and pilots across two kinds of buyer. What it does not yet have is a client willing to be named. Until one is, the case that investment teams will hand a running mandate to an agent rests on Balderton’s judgment and Jha’s record, not on a customer list.