Guide
AI implementation vs. AI training: what operating partners should buy
AI training teaches people how AI tools work. AI implementation sets AI up inside their actual work (their software, templates and daily tasks) and then trains them on that setup. Training changes what people know; implementation changes how the work gets done. For portfolio companies, buy implementation for the teams with repetitive desk work, and use training to support it.
Side by side
| Seats only | Training | Implementation | |
|---|---|---|---|
| What you get | Licenses | Workshops, courses, prompt tips | Workflows set up in the team's tools, then training on them |
| What changes on Monday | Nothing, unless someone experiments | Depends on each person's follow-through | The process itself: templates, drafts and steps are in place |
| How you measure it | Seats bought | Attendance and satisfaction | Active use, hours saved, output |
| Best for | Nothing on its own | Leadership awareness, broad literacy | Teams doing repetitive desk work |
Why training alone tends to fade
A workshop shows people what AI can do in general. Back at their desks, they face the same inbox, the same system and the same deadlines, with a blank chat box and no time to work out where it fits. Some people figure it out; most go back to the old way.
The numbers show the same gap between individual use and company setup. In Liberty Mutual's 2026 Independent Agency Growth Study, 65% of insurance agency staff had used AI for work in the past year, but only 14% said their agency had put an AI tool in place.
What implementation looks like at a portco
- Watch how each team does the work today.
- Set AI up inside the tools they already use, with templates and steps for their real tasks: quotes, proposals, follow-ups, service replies and call notes.
- Train each person on that setup, using their own work.
- Measure use and time saved, and fix what isn't being used.
- Leave behind a playbook, a policy and the numbers.
Where each fits in a portfolio rollout
Use training for the operating partner's team and portco leaders, so they can judge what good looks like. Use implementation for the teams whose work you want to change. In a roll-up, the implementation from the first two portcos becomes the playbook for the rest. See how to roll out AI across portfolio companies.
Questions to ask a provider
- Will you set AI up inside our companies' existing software, or teach general skills?
- Do you work with each team, or run sessions for everyone at once?
- What do you measure, and do you set a baseline first?
- What do you leave behind when you finish?
- Can what you build at one portco be reused at the next?
Questions
Is AI training enough for portfolio companies?
Usually not on its own. Training builds awareness, but if nobody sets AI up inside the team's software and templates, most people drift back to their old process. Training works best as part of an implementation, built on the team's own workflows.
What should an AI implementation leave behind?
Working workflows inside the company's software, trained people, a written AI policy, the numbers that show what changed, and a playbook the next company can use.