Every year, thousands of UK startups raise under the Seed Enterprise Investment Scheme and the Enterprise Investment Scheme. The tax reliefs on offer are extraordinary by international standards, giving investors up to 50% income tax relief on early-stage bets and making SEIS one of the primary mechanisms through which pre-seed capital flows in Britain.
But accessing those reliefs requires navigating a compliance process that has remained almost entirely manual for the past two decades. Before a startup can offer SEIS-qualifying shares to investors, it must obtain advance assurance from HMRC, a formal confirmation that the company and the proposed investment meet the scheme’s eligibility criteria. The application requires detailed information about the company’s structure, its activities, its share classes, and the intended use of funds. It cross-references multiple HMRC guidelines. And it has to be right, because errors do not just delay a round. They can void the tax relief entirely, exposing investors to unexpected tax bills and founders to the reputational damage of a failed compliance process.
In practice, most founders hand this work to a solicitor. The going rate at a mid-tier law firm runs to several thousand pounds. The process takes weeks. And because the underlying guidelines change periodically, the institutional knowledge of what HMRC will and will not accept sits largely inside law firms rather than with founders themselves.
Undo Capital, a London-based platform founded by Valeriia Volkova, has built what it believes is the first AI-native solution to this problem. The approach is straightforward in concept but technically demanding in execution. A founder uploads their pitch deck and financial model. The platform’s AI reads both documents, extracts the relevant information, and uses it to populate the HMRC advance assurance application automatically. It then validates the completed application against current HMRC guidelines, identifies specific problem areas, and flags them for the founder before submission.
The result is a process that previously required a solicitor and several weeks, compressed into something a founder can do themselves in a fraction of the time and at a fraction of the cost.
“The advance assurance process is not intellectually complex,” says Volkova, who spent years working as a product lead in legal technology and co-founded Rattlesnake Group, a London-based design and product studio, before launching Undo Capital. “It is procedurally complex. You need to know which guidelines apply, how HMRC interprets specific edge cases, and how to present the company’s activities in a way that maps cleanly onto the eligibility criteria. That is exactly the kind of work AI is well suited to do, and it is exactly the kind of work that founders should not be paying solicitor rates for.”
The platform goes further than advance assurance. Undo Capital has built an integrated environment that connects cap table management, investor data rooms, document generation and SEIS/EIS compliance in a single workflow. The company is currently developing a feature that will generate a cap table automatically from uploaded documents, extending the same document-to-data logic that underpins the advance assurance tool.
The thinking behind the platform reflects a broader frustration with how founders have historically approached the problem. Mikael Saakyan, Finance and Strategy Lead at Undo Capital, puts it directly.
“For years, startups have adopted a new platform for every new challenge. One system for equity, another for documents, another for investors and another for compliance. The problem is not that founders need more software. They need software that understands how these processes connect. The future belongs to platforms that eliminate fragmentation rather than adding another application to the stack.”
The timing reflects a broader shift in how AI is being applied to compliance-heavy professional services. For years, legaltech and regtech promised to automate document-intensive workflows but largely delivered search tools and template libraries. The generation of AI models capable of reading unstructured documents, extracting structured information, and reasoning against regulatory frameworks has made a qualitatively different kind of product possible.
The broader market makes the opportunity clear. HMRC data shows that EIS and SEIS schemes together have facilitated over £2 billion in investment in a single tax year. As that volume grows, so does the compliance burden on the founders and investors navigating it. The question is not whether the process needs modernising. It is whether AI can do it reliably enough that founders trust it with something as consequential as their investors’ tax relief.
Volkova believes the answer is yes, provided the system is built around the specific regulatory context rather than adapted from general-purpose tools. “The mistake most people make with AI in compliance is treating it as a search engine with a nicer interface,” she says. “What we have built is something that understands the structure of the problem, not just the language around it.”
For UK founders navigating SEIS for the first time, that distinction may turn out to matter quite a lot.
Disclaimer: This article is provided for general informational purposes only and does not constitute legal, tax, financial or investment advice. References to products or services should not be interpreted as endorsements, and readers should seek professional advice where appropriate.