Small business owners are told constantly that they should be "using AI," but far fewer can point to the specific task where it would actually help. That gap between ambition and application is the daily work of Ghost Orchestra, a Tennessee-based AI consulting firm that builds custom assistants for small businesses. The company was recently included in a Peekwire roundup of AI consulting firms in Tennessee, and its write-up there prompted this exchange.
David Wakefield of Ghost Orchestra agreed to answer three questions about how he approaches the earliest conversations with owners, what most AI advice for small businesses overlooks, and the decision owners tend to make too quickly. His answers, lightly edited for clarity, follow.
Start With the Day, Not the Tool
The first question was simple: when an owner says they want to "use AI in the business" but cannot say where it would help, what do you ask first?
Wakefield's answer was direct. He said "I don't start with AI at all. I ask them to walk me through a normal day, from the first phone call in the morning to the last thing they do before they lock up." From there, he asks what the owner finds themselves doing over and over, and what they wish someone else could take off their plate.
That list, he said, is where the answer usually sits. "It's often the repetitive work: typing the same information into two places, chasing quotes, or answering the same questions by email."
For owners trying this exercise themselves, the practical takeaway is to treat a normal day as the source document. Rather than browsing AI tools first, write down the recurring tasks — the ones that repeat daily or weekly and consume time without requiring much judgment. Those are the candidates worth examining. General guidance on identifying repetitive business processes is widely available, but Wakefield's version is deliberately specific: the first phone call through locking up.
The People Question Comes First
The second question asked what comes up in those first conversations that most AI advice for small businesses leaves out.
Wakefield's answer was the people. "A lot of advice starts with the tools, but I think the first thing on a lot of owners' minds is their staff. They'll want to know if this is going to make their people feel replaced, and whether anyone will actually use it."
Ghost Orchestra's stated position is that AI should work as an assistant to the people already there, not as a replacement for them. Wakefield put the practical stakes plainly: "If the staff don't trust it or find it helpful, it doesn't matter how good the tool is."
That framing matters for owners weighing any new system. Adoption is not a secondary concern to be handled after rollout; it is a precondition for the tool doing anything at all. A system that employees avoid, or quietly work around, produces no return regardless of its capabilities. For readers thinking through this, resources on managing technology change with employees can help structure those conversations before a purchase is made.
The Cost of Choosing Too Quickly
The third question was about what owners decide too quickly when they start looking at AI tools, and what that costs them later.
Wakefield pointed to tool selection happening before the job is defined. "It's easy to sign up for something because it looks impressive, and then try to find a use for it afterward." The consequences he described are concrete: paying for software nobody uses, or bending the way the business works to fit the tool.
His alternative is to start with a clear picture of where the time actually goes, then choose one specific job to fix first. That sequencing — problem, then tool — is the through-line of all three answers. It also reflects how Ghost Orchestra describes its own process: a detailed look at how the whole operation runs and what people spend their days on, followed by a custom assistant built around that business.
What Owners Can Take From This
Across the three questions, the same pattern holds. The starting point is observation of the existing business, not a catalog of AI products. The first constraint is the staff who will use the system, not the system's feature list. And the first decision is which single job to fix, not which subscription to buy.
None of this requires a large budget or a technical background. It requires an owner to spend an hour writing down what actually happens in a day, and to ask their team what they would hand off if they could. Wakefield's answers suggest that the businesses most likely to get value from AI are not the ones that move fastest to a tool, but the ones that know precisely which task they are asking it to take on.
For owners in Tennessee and beyond who are early in that process, the Ghost Orchestra site outlines the firm's approach to custom assistants for small businesses. The broader lesson from this conversation stands on its own: define the job before you shop for the tool, and bring your people into the conversation before you bring in the software.