AI rollups promise to bring venture-style technology into fragmented, often overlooked industries. Yet the model depends on something much less glamorous than building software: finding enough suitable businesses to acquire and persuading their owners to engage. 

Speaking with Will Maunder-Taylor on Unsung, Viktoria Izdebska, who founded and exited Austria's fastest growing fintech at 17 and now advises AI rollups across the US and Europe on growth and M&A, explains how AI agents can help buyers identify small business acquisition targets, understand potential sellers and personalise outreach at scale. 

She also talks about where technology creates a genuine advantage and why creativity and human judgement still matter. 

Viktoria had already built companies in fintech and data when she first encountered AI rollups. The concept immediately felt different from other technology trends she had seen.

“The moment I saw the concept of AI rollups the first time, it felt completely different to everything I’ve seen before,” she says. “It felt like it was the next internet boom.”

Traditional rollups create value by acquiring companies in fragmented markets and combining them into a larger group. AI adds another layer by automating labour-intensive processes within the acquired businesses.

This is particularly relevant in sectors such as accounting, property management, managed IT services and HVAC, where many workflows remain manual. Rather than improving margins gradually through scale alone, an AI rollup may be able to change how the underlying work is performed.

Viktoria sees this as a new asset class forming between private equity and venture capital. The model involves buying established companies with customers and revenue, while also making substantial operational changes that carry more risk than a conventional consolidation strategy.

Technology and acquisition capital are only part of the challenge. A rollup aiming to complete 15 or 20 deals a year needs a repeatable way to identify suitable businesses and start conversations with their owners.

AI agents can support this process by researching companies before any contact is made. They can examine websites, company updates and other public information to identify events that may make a business a relevant acquisition target.

The research can then extend to the owner. An agent might consider how long someone has run the company, whether its growth has slowed or whether the owner may be approaching retirement.

These signals cannot prove that someone wants to sell. They can, however, help buyers focus on more relevant prospects and approach them with greater context.

Viktoria is open about sharing how the process works. “I strongly believe in the democratisation of knowledge,” she says. “I don’t think that there’s anything that a person should keep for themselves really.”

That openness reflects a broader point: access to the technology is not, by itself, a lasting advantage. What matters is how thoughtfully the tools are applied.

Detailed research allows buyers to move beyond generic acquisition emails and letters. The message can be adapted to the company, the owner and the circumstances that may be shaping their decisions.

Viktoria says some campaigns have achieved “conversion rates from letters of 10%” at scale. Those results did not come from writing one successful message and repeatedly sending it to the same type of business. They emerged through several rounds of testing.

A campaign might begin with a broad category, such as accounting firms, before narrowing the audience according to age, tenure, company performance and other characteristics. Each batch provides more information about which owners are receptive and which signals may indicate a greater willingness to sell.

The objective is therefore not simply to contact more companies. It is to understand who is most likely to engage, why the timing may be right and how the message should change.

Company information is only one part of the picture. Viktoria also creates personas representing different types of sellers.

One might be an owner who has spent years building a company but has grown tired of administration and wants to focus again on customers and the team. Another may be approaching retirement without realising that their business could attract a buyer.

Viktoria builds an agent around each persona and uses it to explore how that person might respond to different language. A seller who views AI as a threat, for example, may react very differently from someone who sees automation as an opportunity. The framing may need to focus on succession, growth or the possibility of remaining involved after the acquisition.

The method does not require advanced technical knowledge, but it does require structure. “Always use a fresh chat whenever you start something new,” she advises.

She may use one conversation to construct the persona and improve the prompt, then create another dedicated entirely to that profile. The agent remains only as useful as the information provided. Without sufficient context, it may produce generic answers or invent details.

AI makes personalisation easier, but that advantage will not last indefinitely. Growth channels tend to lose their effectiveness once they become widely adopted.

Generic email became crowded, followed by personalised email. Some AI rollups are now using physical machines that imitate handwriting and produce thousands of letters. As more buyers adopt the same approach, its novelty will also begin to fade.

This is already encouraging a return to more personal forms of engagement, including events, dinners, gifts and other tangible ways of creating a connection.

For Viktoria, the winners will not necessarily be those using the most sophisticated automation. They will be those willing to test unfamiliar approaches and search beyond the industries and channels attracting everyone else.

“The winners within the M&A spaces will be the ones that are most creative,” she says. “Look where no one else is looking.”

For more on entrepreneurship through acquisition, small business ownership and Unsung events, subscribe to the Unsung newsletter.