Your team is using AI on the side. Now what?
You probably opened ChatGPT once or twice. You've heard about AI from your peers, your trade publication, your kids. Inside your own company, nobody's said "here's what we're doing about it." And you've started to wonder whether that should be you.
AI raises the floor, not the ceiling. Your most experienced people benefit least.
Here's what's actually true at this stage, before we go any further: you aren't behind. The most common position for small and midsize business owners on AI right now is exactly where you are. The Microsoft and LinkedIn 2024 Work Trend Index found that 60% of leaders say their company lacks a vision and a plan for AI — and that's across all company sizes, not just smaller ones. Two-thirds of executives globally are sitting where you're sitting.
You don't need to become an AI expert. You don't need to invest a quarter's worth of attention. You need a small, specific set of moves you can run in the next two weeks that gets you out of "I'm not sure" and into "I have actual evidence about whether this matters for my company." Five moves below. None of them require permission. None of them cost more than the price of a single AI subscription.
You might be here if…
- You've heard about AI but there's no formal plan. Your peer group is talking about it. Your industry publications are talking about it. Inside your own walls, nobody has said "here's what we're doing about it" — and that's started to bother you, even if you're not sure why yet.
- A handful of employees use ChatGPT; most don't. If you ask around honestly, you'd find one or two people who use it every day, a few who tried it once and didn't see the magic, and most who haven't touched it. There's no pattern to who falls into which group.
- You're not sure if it's worth your attention yet. You read the articles. You feel a little behind. You're also aware you could pour weeks into this and end up with nothing useful. Both feelings are valid; both are pulling in opposite directions.
If most of these don't sound like you, the page you actually want is probably another phase — see the four phases →
Why most owners stall here
The honest reason Phase 1 lasts longer than it should is that the first experience most owners have with AI is mediocre.
You hear the hype. You open a chat window. You type something — maybe a question you'd normally ask a smart assistant. The answer comes back. It's fine. It's not the magic the articles promised. You think "okay, this is another overhyped thing" and you close the tab.
That experience is the single biggest reason small businesses stay in Phase 1, and it has nothing to do with AI's actual capability. There's peer-reviewed research on this. A 2023 study by Brynjolfsson, Li, and Raymond — published in the Quarterly Journal of Economics in 2025 — tracked 5,179 customer support agents through a generative-AI rollout. They found that the productivity gains from AI were dramatically larger for less-experienced workers than for experts. Novice workers saw 34% productivity gains; the most experienced workers saw almost none.
What that means for you, sitting in Phase 1: if you opened ChatGPT and weren't impressed, that's not surprising — you're an experienced operator, and AI helps experienced operators least. The opportunity in your company isn't necessarily for you personally. It's for the people on your team who are doing the repetitive, learn-by-doing parts of the work. They're who AI will help most. They're also probably already using it without telling you.
The other piece of context worth knowing: there's a real gap between official adoption rates and what's actually happening on the ground. The gap between the conservative Census number (9% of small businesses) and the broader US Chamber number (58%) is exactly the gap between "we use AI as a company" and "some of our people use AI on the side." That gap is Phase 1.
What to try this week
Five moves, ordered from tonight to this quarter. The first three cost nothing and you can do them without anyone's permission. The fourth costs the price of a single AI subscription (under $30 a month). The fifth is the deliberate decision you make at the end — yes or no, with evidence behind it.
- 01 free / tonight
Spend 20 minutes using ChatGPT or Claude on a real task
Pick something you'd normally do at the end of your week — drafting a customer email, summarizing a meeting, writing a job posting, putting together a quick analysis from a spreadsheet. Open ChatGPT (chat.openai.com) or Claude (claude.ai) — both have free versions. Paste what you'd normally start with and ask for help.
Don't try to make it perfect. Don't try to learn the "right" way to prompt. Just ask the way you'd ask a smart assistant. The point isn't to become an AI expert tonight. The point is to feel — for yourself — what these tools actually do, where they help, and where they fall short. That direct experience is worth more than ten articles.
Free resource: First 10 Prompts to Try with Your Team (coming soon) - 02 free / this week
Find out who in your company is already using AI
Send a short, anonymous, three-question survey to your team. Question one: "Are you using AI tools (ChatGPT, Claude, Copilot, anything similar) in your work today?" Question two: "If yes, what kinds of tasks?" Question three: "What's stopping you from using it more?"
Most owners are surprised by the answers. The Microsoft Work Trend Index found that 78% of AI users at smaller companies are bringing their own tools to work — without leadership knowing. Your data probably looks similar. The survey is the cheapest way to find out what's actually happening so you can decide what to do about it.
Free resource: 5-Question AI Usage Survey for Your Team (coming soon) - 03 free / this week
Write a one-paragraph acceptable use policy
If your team is using AI on the side, that means there's no rule saying what's OK and what's not. Most owners avoid this because it sounds like a legal project. It isn't. The version your company actually needs is a paragraph: don't paste customer data into AI tools without checking with us, label AI-assisted content if customers see it, ask if you're not sure.
Putting this in writing — even informally, even just in Slack or an email — does two things. It signals leadership has noticed AI exists. And it removes the awkwardness for the employees who've been quietly using AI and worrying they shouldn't be. Both matter more than the policy specifics.
Free resource: One-Paragraph Acceptable Use Policy Template (coming soon) - 04 ~$30/month / this month
Pick one repetitive task and try AI on it for a week
Choose a task that you or one of your people does every week — invoice categorization, customer-email triage, social-media drafting, weekly report generation, anything that follows a pattern. Spend a week doing it the AI-assisted way. Track how long it takes vs. the old way and how the quality compares.
Don't pick a critical or customer-facing task for this. Pick something where a mediocre AI result is recoverable. The goal is honest data — does AI save real time on this real task? — not a heroic transformation. After a week, you'll have a clear picture of one specific use case in your business. That's more than 95% of small business owners have.
Free resource: Buyer's Guide: AI Tools for Small Business Under $30/Month (coming soon) - 05 planning / this quarter
Decide whether AI is officially worth your attention
After the four moves above, you'll have actual data: you've used the tools yourself, you know who's using them in your company, you've put a basic policy in writing, and you've tried one task systematically. That's enough to make the call most Phase 1 leaders avoid: is AI worth a real plan?
If the answer is no, you've made it deliberately, with evidence, and you can move on with a clear conscience. If the answer is yes, you're ready for Phase 2 — which is about what to invest, who to put in charge, and how to get past the disappointment that catches almost every company at that stage. Either decision is the right one for your situation. Avoiding the decision is what keeps companies stuck.
Free resource: Phase 1 → Phase 2 Decision Worksheet (coming soon)
A 40-person commercial cleaning owner runs the four moves over two weeks
Here's the pattern, drawn from the Phase 1 small businesses we see most often. A 40-person commercial cleaning company. Owner-CEO who took over from her father, grew it from 25 to 40 staff, healthy margins on recurring contracts. She has a personal ChatGPT account she got curious about last year. Uses it occasionally to draft emails. Hasn't told her team. Reads the AI articles in her industry magazine on weekends. Feels slightly behind. Hasn't done anything about it.
She runs the four moves above over two weeks. The first night, she spends 20 minutes asking ChatGPT to help her draft a response to a difficult customer complaint — something that would normally take her 40 minutes of staring at the screen. The draft is rough but useful. She edits it down in five minutes and sends it.
The team survey comes back surprising. Her office manager has been using ChatGPT to help draft scheduling communications for three months. Her sales lead has been using it to research prospect companies. Her bookkeeper hasn't tried it once. Two field operations people occasionally use it on their phones. None of them had told her.
She writes a one-paragraph policy in a Slack message: "Use AI tools if they help. Don't paste customer information into them without checking with me first. Label AI-assisted communications if customers will see them. Ask me anything if you're unsure." The reaction is small but positive — three people thanked her for putting it in writing.
By the end of two weeks she has actual evidence. AI is already in her company. It's saving real time for two of her people. There's at least one task — sales prospect research — where she could probably get more value with structured help. She decides Phase 2 is worth the next quarter's attention. The decision is small but it's hers, made deliberately, with evidence.
We don't typically work with companies in Phase 1 — most of you don't need us yet, and the moves above are enough to get you to Phase 2 on your own. If you want to see how engagements work for when you do need help, here's what working with us looks like →
What changes when you move to Phase 2
When you've done the work above and decided AI is worth real attention, you'll start asking different questions. Not "should we?" but "how, and how much?" You'll start spending money — on licenses, maybe on outside help, maybe on training. You'll appoint someone to lead it. You'll run a pilot.
And then you'll start running into a different set of problems. The investment may not pay off as fast as the hype suggests. The pilot will probably teach you things you didn't expect. The good news: the structural reasons most companies stall in Phase 2 are well documented, and the moves to get past them are knowable. We've watched companies move through it; the page on Phase 2 is what we'd tell you next.
That's Phase 2.
Read what to do in AdoptionIf you found this useful, also read:
- Credits & sources — the research backing the claims on this page
- More about Sillewa — who we are and how we work
The other phases
Where you actually are matters. If Phase 1 isn't quite right, here's the rest: