Here is what AI adoption looks like inside most small businesses right now: two people use ChatGPT constantly and tell nobody, three people tried it once and gave up, and the owner is nervous about what customer data is being pasted where.

Licenses don’t fix that. A short, deliberate rollout does. This is the 30-day plan we run for clients from El Paso to Tucson, and you can run most of it yourself.

Week 1: Find out what’s already happening

Before any training, ask every employee two questions, anonymously if you need to: Which AI tools have you used for work in the last month? and What did you use them for?

You will learn three things: who your informal experts are, which tasks people already believe AI helps with, and what data has already left the building. Don’t punish anyone. You asked so you could set rules, not so you could catch people.

Then write the one-page policy. One page, plain English. It needs exactly four sections:

  1. Approved tools and which account to use (business accounts with data-protection terms, never personal ones).
  2. What never goes in: customer financial details, employee records, medical information, anything under an NDA, passwords.
  3. The verification rule: anything AI writes that leaves the company is read by a person first, and numbers are checked against the source.
  4. Who to ask when unsure.

That is the whole policy. Longer ones don’t get read.

Week 2: Train by role, on real work

Generic AI demos produce generic results. Group people by what they actually do and have each person bring three real tasks from their week.

  • Office and admin: email drafting, summarizing threads, turning notes into documents, spreadsheet formulas and cleanup.
  • Sales and estimating: proposal first drafts, researching a prospect, summarizing a spec or RFP, rewriting for a different audience.
  • Field leads and supervisors: voice note to daily log, summarizing a long document on a phone, drafting a safety brief.
  • Owners and managers: analyzing a report, preparing for a hard conversation, drafting a policy, comparing vendor options.

Two hours per group is enough if every minute is hands-on. By the end, each person should have a saved set of prompts they wrote for their own tasks. That personal library is the thing that makes the habit stick.

Week 3: Pick the shared wins

With everyone using the tools, patterns show up. Three people are summarizing the same kind of document. Two are rewriting the same template every day. Those are candidates for a shared workflow: a saved prompt everyone uses, a custom assistant with your templates loaded, or a small automation.

Pick two. Build them simply. Announce them. This is where the time savings jump from individual to company-wide.

Week 4: Measure and decide

Ask the same two questions from Week 1 again, plus one more: How many hours a week do you think this saved you? The answers are rough, but the trend is real. In our experience the honest number for an office role is 3 to 6 hours a week after a month, higher for anyone who writes a lot.

Then decide three things: which tools to keep paying for, which shared workflows deserve real automation, and who your internal AI champion is going to be. Give that person an hour a week to keep things moving.

What goes wrong

  • Training without the policy. People get fast and sloppy. Write the page first.
  • Policy without training. People read the rules as “don’t,” and the informal users go further underground.
  • Teaching tools instead of tasks. Nobody remembers a feature tour. Everyone remembers the afternoon they finished their least favorite report in ten minutes.
  • Skipping the measurement. Without a number, the budget conversation in six months is a guess.

If you’d rather have someone run this with you, our AI Training & Enablement program is this plan with a facilitator, bilingual sessions where needed, and a prompt library built on your documents. We run it on-site across El Paso, Midland–Odessa, New Mexico and Southern Arizona, or live over video. Ask us about it here.