The cost of waiting
At a service firm, routine quotes keep waiting for the founder’s approval. Prospects wait, salespeople chase internally and delivery work gets interrupted; the queue has consequences worth examining before anyone proposes an AI tool.
The first investigation is straightforward: follow several recent requests from arrival to a usable quote. Separate time spent working from time spent waiting. Record why the founder had to become involved and which information was missing.
Perhaps most requests repeat familiar service combinations. Perhaps the difficult cases genuinely require experience. Perhaps the delay begins earlier, with sales collecting an incomplete brief. Those findings point to different solutions.
A business case becomes stronger when it describes a consequence the organization already wants to change, with someone prepared to own it.
A defined problem
Give the problem a clear boundary. Instead of “improve sales with AI,” the firm might investigate preparing standard quote information for an authorized reviewer. That leaves the commercial decision with a person while reducing the work needed to reach it.
Write down the current volume, effort, delay and error patterns. Identify which outcomes matter: a faster response, fewer clarification rounds or more founder capacity for unusual opportunities. Agree how to observe them before selecting a tool.
This also exposes weak assumptions. If the founder delays quotes because delivery has no capacity, faster preparation may not unlock more sales. If every request is different, a rigid template may create more correction work. The diagnosis needs to survive those possibilities.
The other options
Once the problem is clear, compare ways to solve it. Better intake questions could remove missing information. A shared pricing and scope reference could reduce dependence on the founder. A system integration could carry approved details into the quote. AI could help interpret the request and prepare a draft from agreed material.
The eventual solution may include several of these changes. There is no advantage in forcing the whole problem into an AI-shaped answer.
Ask what must be true for each option to work. Is the reference information current? Can the reviewer verify the draft efficiently? Is there a sensible route for exceptions? Would the team use it inside its normal work?
Now planning has something concrete to organize around.
The first investment
Compare this problem with the other candidates using consequences, frequency, feasibility and ownership. A highly visible idea may be less valuable than an unglamorous process that repeatedly delays revenue or creates rework.
Choose a first task small enough to test, with a result important enough to matter. For the services firm, that could mean one common quote type. Use past examples, include awkward cases and measure the total effort required to produce an approved response.
If the trial fails, the business has learned something about a real constraint. If it works, there is a reason to invest further. Either outcome is more useful than an expanding list of possibilities with no accountable owner.
A useful starting point
The next AI meeting can begin with a different question: what problem would we still need to solve if nobody had mentioned AI?
Bring that problem, the people closest to it and a few recent examples. Superfactor can help turn them into a practical decision about process, automation and AI, then design and engineer the improvement the evidence supports.
