AI is rewriting the venture playbook: From roll-ups to founder selection
AI is changing more than the products venture capital invests in. It is widening the types of businesses investors can build and changing how they source and assess the founders behind them. At SuperReturn Venture, Andre Retterath, General Partner at Earlybird, and Roxane Sanguinetti, Founding Partner at Alma Angels, explored two sides of that shift: from AI-enabled roll-ups in traditionally manual industries to the use of technology to identify venture’s next outliers.
• AI could make highly fragmented, traditionally manual service industries more suitable for venture-backed business models.
• Venture roll-ups differ from traditional private equity by using technology to build a scalable core before accelerating growth through acquisitions.
• Large fragmented markets can offer the scale required to generate the outlier returns venture capital depends on.
• AI is also becoming part of how VCs source and assess founders, although investors remain divided over how much weight models should carry.
• Automating parts of sourcing and analysis could allow investors to spend more time building relationships and reaching founders outside traditional networks.
AI brings a different roll-up model to venture
Roll-ups are not new.
Private equity firms have long acquired multiple companies, combined them and used operational improvements, technology and financial engineering to create value.
Retterath believes AI is enabling a different version of that model in venture capital. Instead of assembling several businesses first and improving them afterwards, a venture investor can identify an attractive fragmented market, incubate a technology-enabled core business and then use acquisitions to accelerate growth once that foundation is established.
In the venture world, it’s now really about AI and more of an incubation style.
Andre Retterath, General Partner at Earlybird
The distinction matters because venture capital is looking for a very different return profile.
Turning fragmented industries into venture-scale opportunities
Venture returns follow a power law, Retterath argues: a relatively small number of companies generate a large share of overall performance. That means an investment needs the potential to become an unusually large business.
Fragmented industries can offer an interesting combination. Markets such as accounting and real estate are already substantial, but their fragmented structures have historically made it difficult to build the type of scalable company venture investors typically seek. AI could begin to change those economics.
Today, we have AI, and we can see that with AI, we can suddenly transform many of these traditional industries.
Andre Retterath, General Partner at Earlybird
Service businesses are particularly interesting because many were previously overlooked by venture investors due to their manual nature. If AI can automate more of that work, technology can become a much larger part of the operating model. Acquisitions can then be used to bring fragmented providers onto a more scalable platform.
The roll-up itself is therefore not the innovation. What changes is what technology allows the combined business to become.
From changing companies to changing how investors find them
AI is also beginning to alter an earlier stage of the venture process: deciding which founders deserve attention in the first place. Sanguinetti says investors are already using AI in different forms to source and assess founders.
But there is no consensus yet on where the model should stop and human judgment should begin. During her SuperReturn Venture panel, some investors argued that technology should inform the assessment but never make the final investment decision.
Others were prepared to trust a model even where their own instinct about a founding team pointed in another direction.
We’re definitely seeing that we’re all still learning.
Roxane Sanguinetti, Founding Partner at Alma Angels
Looking beyond credentials
For Sanguinetti, founder selection is not simply about building a better version of the traditional investment scorecard.
She argues that investors need to understand a combination of credentials and underlying characteristics, and determine whether those traits can be identified earlier. Grit and resilience remain familiar qualities for venture investors.
She also highlights a willingness to challenge the status quo as a characteristic associated with exceptional founders. AI and data could potentially help investors assess those traits more systematically.
But the larger opportunity may be to use technology to broaden where investors look in the first place.
Who allocates capital influences who gets funded
That question sits at the centre of Alma Angels’ work. Sanguinetti argues that early-stage financing depends heavily on human networks. Founders need their earliest supporters, and investors naturally form connections with people whose experiences and backgrounds feel familiar. If the population allocating capital is narrow, the founders who gain access to that capital can be narrow too.
Wealth is created on the cap table.
Roxane Sanguinetti, Founding Partner at Alma Angels
Alma Angels therefore focuses not only on women founders but on who controls investment decisions. Its work has expanded from training people to become angel investors and connecting them with women-led businesses to supporting women GPs, family offices, next-generation investors and LPs.
Sanguinetti says the organisation has trained more than 1,000 individuals and family offices since 2020, with those investors backing more than 250 companies. She argues that widening access to capital does not require choosing between diversity and financial performance.
There isn’t a trade-off between investing in diversity and investing for returns.
Roxane Sanguinetti, Founding Partner at Alma Angels
Could automation make venture investing more human?
Perhaps the more interesting implication of AI is that greater automation could create more room for human interaction. Sanguinetti sees technology as a way to reduce time spent on operational work and parts of the analytical process.
The AI tools should be there as tools to essentially save us the time that it takes to do the ops and some parts of the analysis.
Roxane Sanguinetti, Founding Partner at Alma Angels
The purpose of that efficiency is not simply to process more companies.
It can free investors to spend more time meeting founders, building relationships and developing more creative ways to source entrepreneurs outside established networks.
That matters as venture becomes increasingly competitive. More funds and angel investors mean more capital looking for exceptional opportunities. Finding the same companies through the same channels becomes less differentiated.
Expanding venture capital’s opportunity set
Retterath and Sanguinetti approach AI from very different points in the investment process, but their perspectives reveal a similar shift.
For Retterath, technology can expand what venture capital can invest in by changing the economics of industries that were previously too fragmented or manual to scale in a venture-like way.
For Sanguinetti, it can help expand who venture investors find and assess by reducing repetitive work and supporting broader sourcing.
In both cases, AI is widening the opportunity set rather than removing the need for investor judgment. Venture capital will still depend on finding outliers. The question is whether technology allows investors to search for them in places they would previously have overlooked.
