Ashley Abrahams, Head of Technology in
1 October 2026
8-minute read
Artificial intelligence is becoming ubiquitous. It is also frequently wrong. That combination creates both opportunity and risk for private company investors.
At the 2026 Guinness Ventures Investor Conference, I looked at how AI is changing the venture market, what that means when assessing companies, and how we are using the technology within our own investment process.
What is an AI company?
Before looking at the investment implications, it helps to distinguish an AI company from a company that simply uses AI.
“An AI company is one where a model does the work, not one that uses AI to do its work.”
That distinction matters. Building a proprietary model is a very different proposition from incorporating another provider's model into an existing product or workflow.
AI businesses can also create value in different ways. Some reduce the cost of work that previously required many people. Others compress work from days into minutes, allow smaller teams to produce more, identify things a person might miss, reduce risk through continuous monitoring, or make something possible that could not realistically have been done before.
Cheaper
Less labour required
The model performs work that previously required more people.
Faster
Time compressed
Tasks that once took days can sometimes be completed in minutes.
Better
Higher-quality output
Models can identify patterns or risks that might otherwise be missed.
New
New capability
In some cases, AI makes work possible that could not previously be performed economically.
One in six UK venture rounds
AI is no longer a niche within venture capital. The data presented at our conference showed just how quickly it has become a significant part of the UK market.
More AI businesses also means more opportunities competing for capital, more propositions that can look convincing at first glance, and a greater need to distinguish between companies that genuinely have durable advantages and those built on capabilities available to everybody else.
What happens to revenue as AI spreads?
The impact of AI does not fall neatly along sector lines. A more useful question is what happens to a company's revenue as its customers use more AI.
AI can threaten revenue
Defences that may weaken
Hourly billing
AI can reduce the number of hours required to complete a piece of work.
Per-seat software
More productive teams may need fewer users to achieve the same output.
Information advantage
Knowledge that was once scarce may become much easier to access.
AI can support revenue
Defences that may strengthen
System of record
Businesses controlling where important customer data lives may become more embedded.
Proprietary data
Unique datasets that are difficult to recreate can become more valuable.
Usage-based pricing
More AI activity can directly translate into greater customer usage and revenue.
The important point is that being a technology business does not, by itself, determine whether AI is beneficial or threatening. The question is whether greater AI adoption makes the economics of that particular business stronger or weaker.
What this means for investors
For investors considering AI companies through structures such as EIS and VCT, we see four consequences.
01
Cheques can go further
Smaller teams can potentially achieve more milestones with less capital.
02
There is more to choose from
Lower barriers to building software mean more companies can look compelling on paper.
03
Competition affects pricing
The most sought-after opportunities can attract substantial competition between investors.
04
Failure can look different
The risk is not only running out of capital. A proposition may also be replicated far more quickly than in the past.
When assessing an AI company, some of the most useful questions are also the simplest. Whose AI is it? How much of the technology belongs to the company rather than a third-party provider? What data was it trained on? How difficult was it to build? And how long might it take a competitor to reproduce the proposition?
Funding and execution remain important risks. AI adds another question: whether the apparent competitive moat will still exist several years from now.
How we use AI at Guinness Ventures
The other side of the question is how AI can improve the investment process itself. The scale of the opportunity set makes this particularly relevant.
A large part of venture investing is therefore filtering. The challenge is saying no quickly where an opportunity is not a fit, while giving the strongest opportunities enough time and attention.
AI helps us cover more ground. It can support quantitative filtering, gather information, benchmark businesses, interrogate data, flag potential risks and assist with drafting internal material. What it does not do is decide what we invest in.
AI works across the process. So does a person.
Our investment process runs from origination through filtering, review, analysis, evaluation, presentation, investment and portfolio management. AI can assist at every stage, but the division of responsibility matters.
What AI does
✓ Builds sector maps and target lists
✓ Applies quantitative filters
✓ Gathers and benchmarks information
✓ Interrogates data and flags risks
✓ Assists with drafting and portfolio data
What people do
✓ Select the markets worth pursuing
✓ Meet founders and management teams
✓ Decide which questions matter
✓ Judge risks, opportunities and quality
✓ Own the recommendation and numbers
✓ Negotiate terms and sit on boards
Orla: an in-house example
One example is Orla, an in-house tool we have been developing for our SEIS investment process.
It can review a pitch deck, company information and market data before producing one of three possible outcomes for the team to consider.
01
Meet
The company appears sufficiently competitive to warrant meeting the founders.
02
Need more information
The system can prepare a request for additional information for human review and approval.
03
Decline, with reasons
If the opportunity is not competitive against others we are seeing, the system can help explain why.
There is human override at every stage and human approval for anything that goes outside Guinness Ventures. Disagreements between the tool and the team are also logged, helping us understand where the system can improve.
For us, AI is not fundamentally changing what an investment team is responsible for. It is changing how much ground that team can cover.
The benefit is more time spent critically interrogating high-quality opportunities and less time spent on work that can be accelerated safely. The same people remain responsible for evaluating the business, challenging the evidence and ultimately putting their names to the investment decision.
Related resources
Explore more on AI, venture capital and how Guinness Ventures assesses growth companies.
Frequently asked questions
Some of the key questions raised by the growing role of AI in private company investing.
Explore Guinness Ventures
Learn more about how we assess private growth companies, or speak to us if you are building a business that fits our investment approach.
Risk warning: Investments in early-stage and unquoted companies place your capital at risk. The value of an investment may go down as well as up and investors may not get back the full amount invested. Past performance is not a reliable indicator of future results. Tax reliefs depend on individual circumstances and may be subject to change. This article is for information only and does not constitute investment advice.
Sources: Guinness Ventures Investor Conference presentation, 15 September 2026; HSBC Innovation Banking VC Term Sheet Guide 2026.