Short answer. For a mid-market manufacturer, the first 90 days of AI should end with a short list of funded uses tied to real plant or office work, a number for each, and one named owner. Tools and hires come after those decisions, not before.
Days 1 to 30: decide
Walk the work. Talk to the people who quote, schedule, inspect, maintain, and invoice. Write down where time and margin leak today. Then have leadership pick the few uses worth funding and say no to the rest, in writing.
Days 31 to 60: put a number on it and name the owner
For each funded use, agree on what it is worth, what it may cost to try, and the constraint you will not break on the floor. Safety and quality come first. Name one owner per use who can stop it if it is not paying off.
Days 61 to 90: try it, then hire or choose the vendor
With the decisions written, try the first use on real work. When it helps, I build a working prototype or a practical AI tool for your own operations, so the floor can test it before anyone signs a big contract. When it's ready to grow, the company hires onto its own payroll or picks a vendor, and I help it choose. Write any brief from your decisions, not from a sales deck, measure against the number, and report to the board in the same terms.
What usually goes wrong
Buying software before the use is chosen. Forming a committee with no budget owner. Hiring a technical specialist with no defined problem. Each one feels like progress and delays the answer the board wants.
Why listen to me
Career work has focused on commercial leadership, enterprise sales, and sitting with executives while they funded and staffed technical work. For a decade I co-founded and helped grow an applied AI firm to over 100 employees. Before that I led business development for a 250-person technology consultancy and was top-ranked in sales at two Fortune 1000 companies. Clients have spanned small businesses through the Fortune 500.
Engagements are accepted only where consistent with existing contractual restrictions. Dole Consul LLC does not solicit restricted clients, referral sources, employees, or contractors, and does not take work that would compete in a restricted geography or line of business.
Common questions
Where should a manufacturer start with AI?
Start with work your people already do every day where time or margin leaks, such as quoting, scheduling, inspection notes, or maintenance records, and choose the few worth funding.
Do we need a large data project first?
Not usually. Pick the use first, then find out what data it truly needs. Many first uses run on records you already keep.
Who should own AI at a mid-market plant?
One operating leader per funded use, someone who already owns the work being changed.
Do you work with manufacturers outside one region?
Yes, where the work is consistent with existing contractual restrictions. The approach is the same anywhere: lock the decision, the number, and the owner before any hire or vendor.
Keep reading
- What Does a Fractional Innovation Officer Do?
- Fractional Sales Officer or VP of Sales?
- How to Scope an AI Initiative Before You Hire or Buy
- Yes. Number. Owner.
Write me with the company name, what you need, and the constraint. I answer in writing first.
Christopher Dole, Founder, Dole Consul LLC. linkedin.com/in/cdole.