One of the most common topics I’m asked about is how AI-enabled rollups create value post-acquisition. This post attempts to discuss the most common tactics, examples of how they work, and rough success heuristics for founders building in the space.
As a reminder, AI-enabled rollups are quickly becoming the rage among the most talented technologists and founders. The idea is simple in theory: rollup SMBs and services businesses with a combination of equity and debt and leverage AI to drive margin expansion and growth. But very few companies have actually managed to achieve success in practice. Having studied hundreds of these, both the successes and failures, I wanted to share some of the best practices I’ve picked up along the way.
Broadly speaking, there are 3 core levers for value creation in AI rollups (and a surprise 4th one), discussed below, from simplest to most complex:
1. Multiple expansion – multiple expansion is the most straightforward path to creating value and is the playbook traditional PE rollups have executed on for decades now. The idea is fairly simple: buy a bunch of discrete assets at low multiples. Use debt. Integrate those assets together smoothly and do some basic consolidation of back office resources. While this sounds simple, executing M&A is harder than it looks. This is why at Anansi, the first thing I screen founders for and the skillset I think is most important to being successful, is the ability to execute M&A. This is often found in founders who are ex-PE investors or corporate development professionals.
If this basic playbook is executed successfully – the key point being assets can be found and purchased at reasonable multiples for that given sector – you can roughly aim to 2x enterprise value. As an example, in many highly fragmented sectors, you can buy small businesses that do <3m in EBITDA for <5x multiples. By combining at least a few of these entities together, the consolidated entity will often trade at >10x EBITDA and becomes a compelling target for larger private equity firms. The overall returns for lower/middle-market PE rollups, juiced by leverage, have been 20-30% gross IRR. I believe that the following levers can drive even higher returns.
2. Margin expansion – this is where AI expertise starts to become valuable and where sector selection really matters. As a rule of thumb, I try to invest in spaces where at least doubling EBITDA margins is possible. This means existing margins can’t be too high; for example, some IT MSPs often already have 40%+ EBITDA margins, meaning doubling is a tall order. Also, there needs to exist sufficient workflows automatable by software/AI; this is why some physical services like landscaping aren’t necessarily the best sectors either, where physical labor and materials is 75%+ of OpEx.
Margin expansion can be achieved by either automating pieces of the core service and/or automating administrative labor. Often times, the administrative labor piece is lower hanging fruit and can be done by buying off-the-shelf tools versus building proprietary software. As an example, most small business staff a full-time person to answer phones; with the influx of AI voice agents to answer calls, schedule appointments, and answer basic questions, this is a workflow where founders can add a couple points of margin by “taking the VC subsidy” and buying an existing tool instead of building something in-house. The larger jumps in margin more often come from implementing proprietary workflow tools that significantly automate the workflows for say, an accountant or a real estate broker. Regardless, margin expansion is where a lot of founders’ time and effort will be spent in the early days, immediately post-acquisition. My rough heuristic is that the best AI-enabled rollups can 2x EBITDA margins within a year. The very best have even managed to increase margins by 4x+.
3. Top-line growth – this is the lever than I believe is most important, but least discussed. The real prize is to be able to bend the growth curve of these businesses in perpetuity, which is what will enable some of these AI rollups to achieve venture-scale returns. There are many ways this can be done:
- Increasing capacity – E.g. an accountant now has the time to serve 2x the number of customers as AI automates mundane parts of their workflows. Similarly, an insurance broker can now spend all their time prospecting new customers instead of dealing with the admin overhead of pinging customers about renewals, sending emails, etc.
- De novo builds/franchising – after the first acquisition and build-out of a software “operating system,” a company is able to start new franchises, where they get a share of revenue in exchange for giving away their software. Another example is an AI rollup in the primary care space building out de novo clinics as opposed to rolling up existing ones.
- Cross sell/upsell/price optimization – eg. by rolling up insurance brokerages that offer different products (general liability and cyber insurance) you can cross-sell those products. Price increases also are a common tactic that fall into this bucket.
- Scale – by having more scale, companies can often re-negotiate contracts. E.g. clinics at a greater size can often get better reimbursement rates from payors.
- Better sales/marketing – many sharp founders are able to leverage their technical expertise to better market and sell, especially in old school sectors. The most basic example is building out a website, but one interesting example I heard is a founder rolling up roofing companies and leveraging Google Earth data to better target new customers.
The possibilities are endless which is why this lever is so fascinating and important. My general rule of thumb is that if a founder can take a business and roughly double top-line, that’s a great outcome. This essentially means taking a business that’s flat and growing it 15% YoY. The best companies, anecdotally, can actually drive growth to 50%+ YoY.
With these three levers alone, there is tremendous opportunity for value creation. While it is difficult to execute, it doesn’t require much to build a very large and capital efficient business with this playbook.
Let’s walk through a rudimentary example: let’s say a founder finds an interesting sector with high fragmentation, low margins, and lots of opportunity for AI-driven transformation. They can take on 100m in debt to make 20, $5m acquisitions (buying 1m EBITDA businesses at a 5x multiple). They will need to pay back 160m assuming a hypothetical 10% interest loan over 5 years. But if they can double via the first 3 levers respectively (platform multiple goes from 5x to 10x, EBITDA margins go from 10% to 20%, and top-line doubles), you have turned 100m into 800m. And basically you have generated 640m in EV out of thin air ($800m TEV minus $160m in debt). This is the magic of the AI rollup.
Now there’s a final lever here that I believe the best AI-enabled rollups will capitalize on to achieve $10b+ outcomes:
4. Upsell software/new products – I know that this falls into both categories 2 and 3, since it’s a mechanism for margin expansion AND growth, but it feels important enough to be in a category of its own. One of the views I’ve been developing is that the AI rollup/services world will fundamentally change how software is sold and delivered. Historically, companies sold software licenses and upsold services. Salesforce is the canonical example of this. However, I believe we are going to see more and more companies start off by selling a service and then upselling higher-margin software and products, especially as customers “graduate” from their service and want to in-house that function. As an example, an AI rollup in the accounting space can start off by selling basic accounting services, but over time, can either sell their “agent” to a business directly or offer new products that give real-time visibility into inventory, financial metrics, etc. AI rollups are essentially a way to work deeply with a single design partner to build the perfect software for a given vertical.
As always, these ideas are a work in progress and as I continue to learn more from founders in the space, my thinking evolves. I continue to be tremendously excited about AI-enabled rollups and am glad Anansi Capital is all-in on this thesis.