How should an IT services firm roll out AI?

Start where your consultants lose delivery hours, usually proposals, status reports and documentation. Measure a baseline, decide what client data the AI can reach, then train each team on its own work. Guava’s AI Rollout does this in 3–6 weeks for $25,000, on the platform that fits each team.

By Romeo Diaz, founder of Guava · Updated October 30, 2026

The rollout, in order

1. Pick 3–5 workflows Go where the hours are. Proposals and SOWs, status reports, documentation and search across past projects are the usual first four. Rule out what isn’t worth building, such as a chatbot answering clients’ technical questions with nobody checking.
2. Measure a baseline Hours per consultant per week on each workflow, from your PSA, before anyone starts. Without it you can’t show what changed.
3. Get the data ready Decide what each team’s AI can reach, and check client contracts and NDAs for limits on how their data can be used. On Microsoft 365 this is Copilot Readiness: a tenant review, an oversharing and permissions cleanup plan, sensitivity labels and a pilot group.
4. Choose a platform per team Copilot for teams that live in Outlook, Teams and SharePoint. Claude Enterprise or ChatGPT Enterprise for heavy writing and analysis. It’s often more than one.
5. Train on real work Hands-on sessions for your consultants and project managers, using their own proposals, tickets and notes.
6. Track adoption Hours back per consultant per week, measured in your PSA the same way as the baseline. Keep what works and drop what doesn’t.

Why buying licenses isn’t a rollout

Licenses without use cases become shelfware: people try the chat window, the work runs the same as before, and the seats renew anyway. For a firm that sells its people’s time, the test is simple. Either delivery hours came back, or they didn’t.

What it can be worth

The sample AI Assessment for a fictional 60-person IT services firm ranks SOWs and proposals first, then status reports. Together they come to about 28 hours a week back for about $350 a month in AI tools. Those numbers are illustrative, not client results. Your Assessment measures your own.

How Guava does it

AI Assessment $4,000, 2 weeks. Most IT services firms start here. Every workflow ranked Build, Pilot or Don’t build, with estimated hours back and tool costs, and which platform fits each team. Credited in full toward the Rollout.
AI Rollout $25,000, 3–6 weeks, or $21,000 after the Assessment credit. Ready, then Activate (3–5 use cases with measured baselines, configuration and connectors), then Adopt (team training and adoption tracking).
Ongoing Optimization & Governance $4,000 a month, cancel anytime. Adoption, workflow upkeep and AI governance.

Fixed fees, agreed in writing before work starts, and you own everything built. See the engagements

Who you’d be working with

I’m Romeo Diaz: 15+ years of engineering and 5 years building AI agents and workflow automations for clients. Most recently, as Chief AI Officer at an SF-based tech staffing company, I hired and trained 6 engineers who delivered AI projects for its clients. That’s a different kind of firm from yours, and the same problem: getting people who sell their time to work differently. See the track record

Questions IT services owners ask

Which platform should we standardize on?

Maybe none. Teams that live in Microsoft 365 often do best on Copilot; teams doing heavy writing and analysis often prefer Claude Enterprise or ChatGPT Enterprise. The Assessment recommends one per team, and I take no referral fees from any of them.

Do we need the Assessment before the Rollout?

No. If you already know you want Copilot, Claude or ChatGPT, start with the Rollout. If you’re not sure what’s worth doing or which platform fits, the two-week Assessment comes first and its $4,000 is credited in full.

What will AI cost us each month after launch?

It depends on the platform, the number of seats and how much your team uses it. Every scope estimates it up front.

What happens to our data and our clients’ data?

It stays in your accounts, your tenant and your cloud. I use the narrowest access that works, log what I touch and return admin at handover.

Will AI replace our consultants?

No. The goal is to take on more work with the team you have: AI drafts and summarizes, and your consultants review and decide.

Talk it through.

Twenty minutes on where your consultants’ hours go. Prefer email? romeo@workwithguava.com

Book a 20-minute call