The 5 steps at a glance
- Find where your team’s hours actually go.
- Pick one task, not ten.
- Give your team one safe, company AI tool.
- Write a one-page AI use policy.
- Automate that one task end to end, then measure.
The rest of this guide walks through each step, what to skip, and when it makes sense to get help.
Why does starting with AI feel so overwhelming?
Because the noise is real. There are thousands of AI tools, a new “must-have” every week, and plenty of people selling courses on keeping up. Most of it is aimed at someone with more spare time than a business owner has.
The good news: most businesses with 10 to 100 people are still early. According to the U.S. Chamber of Commerce, 58% of small businesses reported using generative AI in 2025, up from 23% in 2023, and many of those are still experimenting rather than getting consistent results. You haven’t missed the moment. You just need a calmer way in.
Step 1: Find where your team’s hours actually go
Don’t start with tools. Start with time. Ask your team three questions:
- What do you copy and paste? Moving information between email, spreadsheets, your CRM and your accounting tool is classic automation territory.
- What do you chase? Invoices, signatures, documents, replies. Chasing is predictable and repetitive.
- What do you do every Monday? Weekly reports, status updates and schedules are easy wins.
Write the answers in a simple table:
| Task | Who does it | Hours per week | How annoying (1 to 5) |
|---|---|---|---|
| Replying to new quote requests | Office manager | 6 | 4 |
| Entering supplier invoices | Bookkeeper | 5 | 5 |
| Monday sales report | Owner | 2 | 3 |
Half an hour of honest answers usually shows you more than a month of reading about AI.
Step 2: Pick one task, not ten
Rank the list by hours per week and how easy it is to fix. The best first task is high-hours, repetitive, and low-risk if something goes wrong. Replying to new leads, sorting email and processing paperwork are common first picks.
What to skip for now
- Custom AI apps. Rarely needed for a business your size. Off-the-shelf tools plus a few automations cover most needs.
- Chatbots nobody asked for. A website bot that annoys customers costs more than it saves.
- Tools that duplicate what you have. Many AI features are already built into software you pay for.
Step 3: Give your team one safe, company AI tool
If people are using AI already, they’re probably doing it on personal, free accounts. That’s the biggest risk most businesses don’t see: company information pasted into tools the business doesn’t control.
Pick one AI assistant and put everyone on a business plan:
- ChatGPT (business plans): a strong all-rounder that many people already know.
- Claude (team plans): strong at writing and long documents.
- Microsoft Copilot: the natural fit if your business runs on Microsoft 365.
Not sure which? Read best AI tools for small business.
Step 4: Write a one-page AI use policy
Keep it to one page. Your team will actually read it. Cover:
- What’s OK: drafting emails, summarizing public or internal non-sensitive documents, brainstorming, editing.
- What’s not: customer financial data, health information, passwords, anything under a confidentiality agreement, unless an approved tool is set up for it.
- Which tool to use: the company account, never a personal one.
- Who to ask: one named person for questions.
- Check the output: AI can be confidently wrong. A person reviews anything that goes to a customer.
Step 5: Automate one task end to end, then measure
What is an AI automation?
An AI automation is a set of steps software does for you automatically. For example: a new quote request arrives, AI reads it, drafts a helpful reply, adds the lead to your CRM and alerts your team, all within minutes.
Automate the task you picked in Step 2 from start to finish. Half-automated tasks often create more work, because someone still has to stitch the pieces together.
How to measure hours saved
- Before you build, time the task for a week (Step 1 gives you a starting point).
- After it’s live, time what’s left: checking, fixing, exceptions.
- The difference is your real saving. Then pick the next task.
Should you do it yourself or hire help?
Doing it yourself costs time. It works well if someone on your team enjoys this kind of work and has a few hours a week to spare.
Hiring help costs money. It makes sense when your time is better spent running the business, when several systems need connecting, or when data safety matters and you want it set up properly the first time.
A middle path many owners take: get a written plan first, then decide what to build in-house and what to hand off. That’s exactly what Spheric’s AI Game Plan is for. If you’re curious about budgets, read how much AI consulting costs.
Where do you stand today?
Take the free AI readiness assessment. Ten questions, about two minutes, and you’ll get three next steps for a business your size.
