Practical guidance for founder-led businesses on mergers and acquisitions, valuation, deal structure, and the realities of implementing AI in a lower-middle-market company.
First-hand accounts of building and operating AI systems inside a working advisory firm — what pays back, what breaks, and how to judge a vendor.
We run one. It answers inbound calls, works out what the caller wants, and books meetings into the calendar without anyone touching it. Here is an honest account of where that works and where it does not.
The first project determines whether there is a second one. Choose badly and the program stops there, regardless of how good the technology was.
If your AI project has been nearly finished for four months, this article is about you. It is also about us, because we spent months in exactly that state.
The market is noisy and the barrier to claiming expertise is low. Here are the questions that separate people who have deployed systems from people who have read about them.
This is the least interesting thing we automated and the clearest payback. That combination is not a coincidence.
If your business holds anything sensitive — financial records, client details, anything covered by an agreement — connecting an AI system to it deserves more thought than it usually gets. Here is the short version of what matters.
Most businesses that say they want an AI agent want an automation, and would be better served by one. The three things get sold interchangeably and they are not interchangeable at all.
Almost nobody publishes real numbers on this, which is why the question keeps getting asked. Here is a straight answer, with the caveat that costs in this market move quickly.
Ask ten vendors what an AI agent is and you will get ten answers, most of them designed to make whatever they are selling sound like one. Here is a definition that is useful if you run a business rather than build software.
The honest answer is: rarely a whole role, often meaningful parts of several. Which sounds like a dodge until you look at how roles are actually composed.
How lower-middle-market businesses are valued, how transactions are structured, and what a sale process actually involves.
Technology rarely decides whether a deal happens. It regularly decides how much is held back at closing.
Most owners hear a multiple before they understand the number it applies to. That is the wrong way round.
A well-run sale takes 6–9 months from engagement to closing. The preparation that determines the outcome takes 12–24 months and happens before any of that starts.
The most common reason an AI deployment underperforms is not the model. It is that nobody told the system how the business actually works.
General advice about business automation is nearly useless, because the work that repeats in a car wash looks nothing like the work that repeats in a collision shop.
Two offers at the same headline price can leave very different amounts in your account.
Most failed business sales do not fail at the negotiating table. They fail months earlier, in decisions that looked reasonable at the time.
Most businesses that believe they have an AI problem have a subscription problem.
Most owners think about price long before they think about who is paying it. That is the wrong order, because the buyer type determines the price, the structure, how long the process takes, how…
Two projects sit on most owners' lists and never get done. One is documenting how the business runs. The other is automating some of it.
The five long-form sector reports are indexed on industry research, the two AI pillar articles and everything else on the resources page. Short answers to common questions are on the FAQ.