A small home service company builds an AI playbook by naming one owner for AI, writing a one-page set of rules, picking two or three jobs where AI saves real time, and reviewing results every week for a month. You do not need a tech team. You need a clear owner, simple rules and the discipline to measure what changes.
Key takeaways
- MIT’s 2025 GenAI Divide report found 95% of companies saw no measurable profit from their AI pilots.
- Boston Consulting Group’s 10-20-70 rule says 70% of AI success depends on people and processes.
- Katie King’s AI Playbook method uses a small group of champions meeting weekly, which a contractor can shrink to one owner and a few staff.
Most contractors start with AI the same way: someone signs up for a tool, tries it twice, and it fades. Nothing was wrong with the tool. Nobody owned it, nobody decided what it was for, and nobody checked whether it helped. A playbook fixes all three.
Before you add AI inside the business, see how AI already describes it with the free 60-second AI visibility check.
What is an AI playbook?
An AI playbook is a short, living document that says which AI tools your company uses, what they are used for, what data they may never touch and who checks the results. Katie King, author of AI Strategy for Sales and Marketing, builds playbooks with cross-department “AI champions” who meet weekly and produce a first version in about six meetings.
For a company with five to fifty employees, the same idea works with fewer people: an owner or office manager, one lead technician and whoever handles the phones.
Why do most AI projects fail?
Most AI projects fail because they add a tool without changing how work gets done. MIT’s 2025 report found 95% of company AI pilots produced no measurable profit, and that companies often spent the most on sales and marketing while back-office uses like admin and scheduling paid back better. McKinsey found redesigning workflows was the biggest driver of AI value.
Harvard Business School research with Boston Consulting Group adds a warning. AI made consultants much faster and better on tasks it handled well, but on tasks outside its abilities, they were 19 percentage points less likely to get the right answer. AI is uneven. Your playbook decides where you trust it and where a person checks.
What should be in a contractor’s AI playbook?
Keep it to one page. It should answer who owns AI, what it is for this quarter, what it may never do, and how you know it worked. Review it every month and change it as you learn.
| Section | Example for a 12-person HVAC company |
|---|---|
| Owner | Office manager, 2 hours a week |
| This quarter’s goal | Cut missed calls and speed up estimate follow-up |
| Approved tools | One chat assistant, the CRM’s built-in AI, call answering |
| Never put into AI | Payment details, alarm and gate codes, health information |
| Human check required | Anything a customer reads, any price, any promise |
| Measure | Missed calls per week, estimate close rate, hours saved |
Book a free call with Mike to sketch your one-page AI playbook in 30 minutes.
Where should a contractor use AI first?
Start where time leaks and mistakes are low-risk. The best first uses for contractors are repetitive writing and answering tasks that a person can quickly check. Avoid letting AI set prices, diagnose equipment or make promises on its own until you trust it.
- Missed-call and after-hours answering that books into your calendar.
- Estimate follow-ups drafted from your notes and sent after a person reviews them.
- Review replies that thank customers by job and town, checked before posting.
- Job write-ups turned into short website case studies and Google posts.
- Office admin like summarizing calls, cleaning customer records and drafting schedules.
What does a 30-day AI rollout look like?
A 30-day rollout runs in four weekly steps: pick, test, adjust and decide. At the end of the month, keep what clearly saved time or won jobs and drop the rest. That rhythm prevents the most common failure, which is paying for tools nobody uses.
- Week 1: Name the owner, write the one-page playbook and pick two uses.
- Week 2: Run both uses on real work and log time saved and mistakes.
- Week 3: Fix the prompts, rules and handoffs that broke.
- Week 4: Compare results to your goal, then keep, change or cut.
Frequently asked questions
Does a small contractor really need an AI policy?
Yes, a simple one. Without rules, employees paste customer details into random tools, publish unchecked AI text and pay for overlapping apps. A one-page playbook sets which tools are approved, what data stays out and who checks the work. It takes an hour to write and saves months of mess.
What AI tools should a home service company use?
Start with tools you already have: the AI built into your CRM, scheduling or phone system, plus one general chat assistant. Add specialized tools, like an AI receptionist, only when a clear problem justifies them. The right tool matters less than having a clear use and a person who owns it.
Who should own AI at a small company?
Usually the office manager or operations lead, not the owner and not the newest hire. They see where time is wasted, talk to customers and can enforce rules. Give them a couple of hours a week, a clear goal for the quarter and the authority to drop tools that do not help.
Can AI replace my office staff?
It is far more likely to make them faster. Studies of AI in customer support found productivity gains of about 14% on average and larger gains for newer staff. In home services, AI works best handling repetitive tasks so your team spends more time with customers and less on admin.
The bottom line
AI success in a small company is mostly about people, rules and follow-through. Name an owner, write one page, start with two uses and measure for 30 days. Book a free call with Mike to build yours.
Sources
- MIT Project NANDA, The GenAI Divide (2025), via Virtualization Review
- Boston Consulting Group, Closing the AI Impact Gap (2025)
- McKinsey, The State of AI
- Dell’Acqua et al., Navigating the Jagged Technological Frontier, Harvard Business School
- Bigeye, interview with Katie King (2026)
Keep reading
