Site icon Kevin Waites Consulting Services

Tried AI and it didn’t stick? Onboard it like a new hire

When a leader tells me, “We tried AI here, but it didn’t stick,” I always ask the same question. How did you onboard it?

Most of the time the answer goes something like this. They bought a few chatbot licences, someone posted a link in Slack, and everyone hoped the team would get more productive on its own.

What happens next is pretty predictable. The most tech-savvy person on staff uses it for everything, the overwhelmed coordinator never opens it, and the organization as a whole gets almost nothing out of it.

If AI is supposed to add capacity, it can’t be an optional browser tab. I’d onboard it the same way you’d onboard a new teammate. When a new coordinator joins a small team, you walk them through your culture, hand them the brand guidelines, explain the quirks of your donors and show them what’s theirs to handle. AI needs that same welcome, plus a clear plan for how work passes between your people and the AI.

Set the shared rules once, for everyone

How do you make sure every AI tool across your organization uses the same voice, follows the same privacy rules and knows the same background?

Asking every staff member to paste a long list of rules into their prompts every morning won’t work. Someone will forget, usually on a busy day.

The better way is to build it into the setup, in two layers:

Once the base rules live in the setup, your safety doesn’t depend on someone remembering to prompt carefully. Consistency becomes part of how the system is built, so you don’t need a training seminar to get it.

How to onboard AI like a new hire

  1. Write the welcome package. Pull together what you’d give a new coordinator on day one: your brand guide, privacy rules, a short background on your programs and donors, and a glossary of the terms your team uses.
  2. Put the org-wide rules where every tool reads them. Most AI tools have some kind of organization or project-level instructions. Put the non-negotiables there so they don’t depend on each person’s prompt.
  3. Give each department its own notes. Ask each team lead to write a page on how their area works, and to keep it current.
  4. Map the handoffs. For each role, list which tasks go to the AI, what comes back to a person and who reviews it before anything goes out.
  5. Hold a one-month check-in. Treat it like a probation review. Talk about what’s working, where people are working around it and which rules need updating.

For leaders: change what you measure

If you listen closely to why a stretched nonprofit team pushes back on AI, there’s often a quiet worry underneath. “If this tool handles my reporting, my data cleanup and my drafting, what am I supposed to do all day?”

That’s a completely rational fear. The headlines are full of AI replacing jobs, so nobody’s going to rush to adopt a system they think is coming for theirs.

You can take that worry off the table on day one by changing what you measure. Nobody on your team was hired to do data entry. You hired them for their judgment and for how much they care about the people you serve. The goal is to automate the overhead that keeps a program manager at their desk until 7 p.m., so they can get out into the community and do the work only a person can do.

So when you introduce AI, pitch it to your team as time back. To most staff, “efficiency” sounds like a cost-cutting word.

Questions worth asking before you roll anything out:

Organization-wide rules that load automatically for every agent are part of how I’ve set up Cosly, the AI staff team I’m building for small nonprofits.

Exit mobile version