When taking over a family business, bring AI in by starting with one process that costs real time and whose improvement would be visible to everyone, including the outgoing generation. Before touching any system, have the conversation about what AI will protect rather than what it might displace. Who helps: a specialist who works with smaller businesses and can explain results to a skeptical board, not a firm that sells enterprise platforms.
Why is the succession moment the right time to raise AI?
Taking over a family business brings two things at once: more permission than usual to question how things work, and less runway than usual to get it wrong. Most family businesses have already moved past debating whether AI is relevant. PwC's 2025 Global Family Business Survey found that 61% of family businesses see AI experimentation as a growth opportunity. KPMG's 2026 Global Family Business Report found that 44% of family businesses are actively piloting AI use cases and 20% have already implemented it at scale.
That momentum creates its own opening. KPMG's 2026 report also found that 38% of surveyed businesses have no AI governance framework in place. AI strategy and deployment is the largest capability gap, cited by 43% of respondents. Adoption has outrun governance. An incoming leader who can frame that part of the work is more useful to the business, and less threatening to the outgoing generation, than one who arrives to announce a new initiative.
The risk matches the opportunity: you are untested. Moving too fast on a system the outgoing owner cares about is the quickest way to lose the collaboration you need. The frame that works is protecting what was built, not replacing it.
Which process should you start with?
Pick the process that is already causing visible pain. A good first process has a cost that is felt every week, an output that is easy to judge without technical explanation, and an improvement that would be obvious to anyone in the room. Scheduling, quote generation, order tracking, and invoice matching tend to fit. So does any process where the outgoing owner can name the single person who carries the whole thing in their head, because that concentration of risk makes the argument without you needing to explain what AI is.
Before any system is touched, writing down exactly how the process works today is the step that makes a pilot run faster and fail more informatively. A process that lives in someone's head cannot be handed to an AI system.
Say the business receives supplier invoices by email and someone matches each one to a purchase order by hand. That person works through thirty invoices a day, each check taking around four minutes. An AI tool can match invoices automatically and surface only the ones with discrepancies, replacing a stack of routine confirmations with a short list of exceptions. The outgoing owner can see exactly what is automated and what is still reviewed by a person.
How do you bring the previous generation along?
The outgoing owner built the business on judgment calls that were correct enough, often enough, to last. A new leader arriving with a technology agenda can read as a correction of those calls, even when that is not the intention.
KPMG's 2026 Global Family Business Report found that AI strategy and deployment is the largest capability gap identified by family businesses, cited by 43% of respondents, 13 points ahead of the next-ranked gap. The gap is in sequencing and governance. The training conversation lands only after a specific use case has been identified; proposed before that, it feels like a solution looking for a problem.
The sequence that works: pick the use case first, frame it in terms the outgoing owner already measures, show what the output looks like before anything goes live, and ask for approval before the pilot starts. The most common concerns are errors, loss of control, and staff resentment. Each has a practical answer, but only if you have asked the question first.
How do you get your team on board?
Staff adopt tools that make their own days easier. How a team makes the shift follows a social pattern, not a rollout schedule.
Find the one person in the team who is already curious. Give that person the tool first, with enough time to develop a real opinion on it. They will explain the tool to colleagues in language that makes sense to the people they work alongside, which a new owner cannot yet do. The second adopter follows the first one, not the announcement from leadership.
Teams rarely move at the same pace. The gap between the curious and the skeptical is a sequence to follow. Running everyone through a group training on day one tends to produce polite attention from some and visible disengagement from others. The same session three months later, after a few people have formed a real view on the tool, lands differently.
What does working with a specialist actually look like?
The right question is who will help you get the first project right, not who will transform the business. A specialist who works with smaller businesses knows how to build a case for an owner-operated board, build something with a small team, and leave something the team can maintain without ongoing support.
A typical first engagement for a business in transition runs four to eight weeks: a workflow audit to identify the right starting point, a scoped pilot on one process, a working output the team owns, and a recommendation on what to try next. The engagement ends with something running, not with a report.
The left column below shows what a solo first attempt typically looks like; the right is a specialist-guided pilot.
| Process selection | Chosen by intuition; higher-value candidates often missed | Audit identifies the starting point with the clearest return |
| Speed to first result | Months of tool learning before anything is in production | Four to eight weeks to a working pilot |
| Case to the outgoing generation | New leader explains it alone, with limited track record | Specialist frames the business case with data from comparable projects |
| What you own at the end | A tool you learned on the job; hard to hand off to the team | A process the team runs independently |
Deciding whether to bring in an outside partner comes down to whether they can show something working in a business at your scale. The questions that matter in a first conversation: which process are we starting with, how will we know if it is working, and what happens if it does not.
Frequently asked questions
Should I tell the outgoing generation about my AI plans before or after scoping the project?
Before. Arriving with a fully formed plan reads as a decision already made. Arriving with a question, such as asking what the outgoing owner knows about the process you have in mind, opens a conversation and often surfaces history you do not have yet. The outgoing owner may know why a previous attempt at change failed or which staff members will need more time.
What if the previous owner is actively against AI adoption?
The strongest argument is competitive rather than generational. KPMG's 2026 Global Family Business Report found that 64% of family businesses are already actively deploying AI. The conversation in the sector has moved past whether AI belongs in the business. That argument lands better when it comes from external research than from the incoming leader personally. A short look at which tools competitors in your sector are running in production can shift the conversation from a values debate to a practical one.
How do I tell whether an AI consultant has real experience with businesses like mine?
Ask for a conversation about a project they ran at a company your size, not the headline client on their credential slide. The details tell you more than the name: which process they started with, what resistance they encountered, and how they measured the result. A specialist who cannot describe the specifics of a past engagement probably does not have the relevant experience.
What if the first AI project does not produce results?
A well-scoped pilot is designed to give a clear answer, not guarantee a success. A pilot that fails on a narrow, reversible test still tells you something useful. The typical failure modes are that the process was not documented well enough before the pilot started, or that the output was too hard for the team to verify on their own. Both are fixable before a second attempt, and neither requires writing off the approach entirely.
Is there a risk that pushing for AI makes me look like I am changing too much too fast?
That risk is real and is managed by which process you choose first. An AI project that improves something the outgoing owner already cares about reads as understanding the business, not disrupting it. The risk lives in choosing a first project that is visible to clients before it has been tested, or that touches a process the previous generation sees as central to the company's identity. Start with something internal and measurable.
Taking over a family business and trying to get the first AI project right are two challenges that run in parallel. Bring us the process that costs your team the most time each week; we will work out whether AI is the right fix and what a pilot would look like.