Rolling Out Chat GPT Isn’t AI Strategy

By Chris Morbitzer, CEO of NorthBuilt and presenter at FEMA’s Marketing & Distribution Convention.
One shortline manufacturer I worked with this fall already had ChatGPT, but their website listed over 1,000 dealer locations whose information had gone stale because updating them by hand would have taken nearly two weeks of one person’s time. Together we pointed their ChatGPT account at that one job. Now AI reads each dealer’s public website and proposes updates, and a person reviews them in batches before anything changes.
That’s the difference between having AI and using it. Giving everyone on your team a ChatGPT license isn’t the same as adopting AI. It’s buying access. Without a purpose, training, and a way to measure results, access alone changes very little.
Here’s what needs to happen first.
1. Start with one task, not one tool.
Pick one task your team repeats every week, like quoting, warranty claims, or cleaning up field data, and aim the tool at that.
2. Train people on that task.
Handing out ChatGPT licenses without training is a recipe for uneven results and silent misuse. Train people on work they already know well, so they can tell when the output is wrong.
3. Set some ground rules.
Governance sounds heavy, but it can fit on one page. Make it clear to your team which accounts to use and not to use, what data stays out completely, who reviews anything before it reaches a dealer or customer, and who owns the result.
4. Pick one number before you start.
A 2025 MIT report found that 95% of the companies it studied got zero return from their generative AI pilots. Pick a metric—hours saved, turnaround time, or errors caught. If you can’t tie the result to dollars or a real business priority, you can’t tell whether or not it worked.
5. Give it 90 days, then decide.
A short pilot beats a company-wide rollout. At the end, check your number, then expand it, adjust it, or drop it.
When I talk with manufacturers about AI, I often hear “we’ve got to use AI” said in the same way as “I’ve got to get to the gym.” Why are you going to the gym? What are you going to do once you get there? AI is like that. It has no value without a specific purpose in mind.
Before you hand out another license, name the task it’s for and the number that will tell you it worked. Then you’ll know where AI is paying off, and where it’s just another subscription.
Chris Morbitzer is CEO of NorthBuilt, FEMA’s Preferred Technology Partner. NorthBuilt is a Midwest software and technology advisory firm that works with independent and family-owned equipment manufacturers and their dealers to align their technology with their business goals and priorities. FEMA members can book a complimentary AI Working Session at femaaiworkshop.com or reach Chris directly at chris.morbitzer@northbuilt.com.
NorthBuilt.com | FEMA Member Since 2025

