Five lessons on building an AI business investors can believe in

TL;DR

An investable AI startup needs more than fast, affordable technology. Investors look for a defensible advantage, evidence of customer demand and a strong founding team. Founders should use AI to validate ideas more quickly, prepare for investment before they need it and choose their investors carefully.

Lesson 1: AI is not a moat

The technology itself is becoming increasingly accessible. Aleksa argued that lasting advantage will come from what surrounds it: proprietary data, specialist expertise and strong customer relationships.

Founders should also think carefully about how they reach customers. If a business becomes an invisible step in somebody else’s journey, it can easily be replaced.

“If AI became ten times better and ten times cheaper tomorrow, would your business get stronger – or easier to copy?”

Lesson 2: Use AI to learn faster, not simply build more

AI has dramatically reduced the cost and time involved in developing software. The danger is using that speed to add features nobody has asked for.

Aleksa urged founders to build something small, put it in front of real users and gather evidence. The important question is no longer whether something can be built, but whether it should be.

“Ten-times-faster engineering should mean ten-times-faster validation – not ten times more features.”

Lesson 3: Answer three questions clearly

For Elliot Smith, CEO of EHE Venture Studio, an investment proposition must quickly answer three questions: what, why now and why you?

That means showing that the business addresses a problem customers will pay to solve. Investors will also examine the market opportunity, the timing and whether the founding team has the expertise to deliver.

“There’s got to be substance behind the deck,” he said. “Probably the most important thing for me at pre-seed and seed is the quality of the founding team.”

Lesson 4: Prepare before you need investment

Dr Sheela Sharma, CEO of Portobello Business Centre, challenged the assumption that every startup needs outside funding from day one. Reaching the proof-of-concept stage with limited resources can demonstrate how effectively a founder will use an investor’s money.

However, founders should still prepare early. Corporate records, financial information and the data room should be kept investment-ready rather than assembled in a hurry when an opportunity appears.

“Investors will respect the individuals who have got as far as possible with as little as possible,” she said.

Preparation also extends to tax and regulation. Gareth Hewins of Philip Hare & Associates warned founders to build sufficient time into the fundraising process for SEIS or EIS advance assurance. Charles Kerrigan, Partner at CMS, advised AI businesses to maintain an inventory of the models they use, classify the associated risks and keep proper records.

Lesson 5: Choose your investors carefully

Investment is not a one-way assessment. Sheela warned that founders can become so focused on securing capital that they fail to consider whether an investor is right for the business.

“You should be assessing them as much as they are assessing you,” she said. “Don’t think the choice is in one direction only.”

There are some things AI can't replace

The arrival of AI may have changed the economics of building a company, but the fundamentals remain familiar. Successful founders still need to solve a real problem, understand their customers and prove that people will buy what they are creating. The technology can accelerate that work; it cannot replace it.

Sources

• EHE Ventures (S)EIS AI Growth Fund
• GOV.UK: Tax relief for investors using venture capital schemes
• GOV.UK: Using venture capital schemes to raise money
• European Commission: The EU AI Act
• Workspace: Salisbury House