Ask "what should we do with AI" and the answers come from people who sell AI
An owner who asks "what should we do with AI" gets answers from people who sell AI. Start from a different list.
Write down the five jobs in the business that eat the most hours each week and that someone complains about. Typing in purchase orders that arrive as PDFs. Chasing unpaid invoices. Rebuilding the same schedule every Monday. Answering the same eight customer questions by phone.
Next to each, write what it costs: hours times wage, plus the errors, plus whatever slips because those hours are gone.
The AI surveys I read this year start after a company has already picked a project. They say almost nothing about how anyone decided what was worth doing, and that is where small businesses get stuck. Your staff know where the friction is. You know what matters to the business. Neither of you necessarily knows which of those jobs software can now take on, or at what cost.
Some of the five may suit AI. Others may be solved by ordinary software that has been around for years, or by deciding the work no longer needs doing. You only find out by looking at the job first.
Then take the list to the next demo and ask the vendor which line their product moves, and by how much.