An AI champion is someone already inside your business who owns how AI gets used day to day. They know your customers, your people and your processes, and they pair that knowledge with tools like Claude or ChatGPT to cut admin, find sales and remove operational drag. For most small businesses, an internal AI champion beats handing AI to an outside agency.
- 80% of small and medium business staff who use AI bring their own tools to work, and 61% of their leaders say there is no plan (Microsoft, 2024).
- MIT’s 2025 GenAI Divide report blamed most failed AI pilots on a learning gap in how businesses fit AI to their work, not on model quality.
- Pick a trusted person who knows how the work really gets done and that others already ask for help.
- Start with four jobs: customer service analysis, ad and website analysis, sales notes, and a daily report loop.
Most of the conversations I have with business owners right now start the same way. They know they should be “doing AI”. They are not sure who should do it. The choice usually looks like two options: pay an outside agency, or do it from within. My answer is to do it from within, and to give the job to a named person. That person is your AI champion.
What is an AI champion?
An AI champion is an existing team member who is given the job of making AI useful across the business. They are usually the person who already understands your people, your culture and how things actually get done day to day, and who injects that knowledge into AI tools to work faster and remove friction.
Banks and consulting firms run this as a formal program with thousands of champions (Citi’s network passed 4,000 people, according to Lead with AI’s guide to champion programs). Microsoft gives small businesses the same advice in plainer terms: find the people who are enthusiastic about AI, form a small group, and make them your champions.
In a business of 5 to 50 people, you don’t need a network. You need one person, with protected time and a clear brief.
Why does AI adoption stall without an AI champion?
AI adoption stalls because nobody turns it into a habit. Someone buys a subscription, sends a link, maybe runs one lunch-and-learn, and within a few weeks usage drops back to a couple of keen people. Small businesses don’t have a training department or an internal comms team to keep nudging, so it has to be one person’s job. As BizTech put it in a 2026 piece on champion networks, licensing a tool and sending out training videos is not enough to change how people work.
The numbers back this up. Microsoft and LinkedIn’s 2024 Work Trend Index found 80% of AI users at small and medium businesses were bringing their own AI tools to work, while 61% of their leaders said the company had no plan. So your team is probably using AI already, with no shared prompts and no rules about customer data.
When MIT’s NANDA initiative reported in 2025 that around 95% of the generative AI pilots it studied delivered little to no measurable profit impact, the lead author blamed a learning gap rather than the quality of the models: generic tools that never get fitted to how the business actually works. Fitting AI to how your business works is the AI champion’s job.
Should you use an AI champion or an outside agency?
For most small businesses, an internal AI champion is the better first move, because the hard part of AI is knowing what to ask, and your own people already know. An agency can set up tools quickly, but it doesn’t know which customer complaint costs you the most, which quote template your estimator hates, or why a job went sideways last month. Your champion does.
| Internal AI champion | Outside agency | Owner does it alone | |
|---|---|---|---|
| Knows your customers and processes | Yes, from day one | Has to learn them, at your cost | Yes |
| Trusted by the team | Yes, they are a peer | Often seen as outsiders | Depends; can feel top-down |
| Ongoing cost | A few hours of an existing wage each week | Monthly retainer or project fees | The owner’s time, the scarcest thing you have |
| Knowledge stays in the business | Yes | Leaves when the contract ends | Lives in one head |
| Speed to first result | Slower without training | Fast for setup | Slow, competes with everything else |
The weak spot of the internal option is the last row. A champion without training spends months working out what an outside expert could show them in a week. That is why the best setup is usually an internal champion with outside training, rather than one or the other.
Who should be your AI champion?
Your AI champion should be the person others already go to for help. The best champions are trusted and good at explaining things, and they don’t need to be technical. They often come from operations, admin, customer service or sales, because those are the people who see where time actually leaks.
What does an AI champion actually do?
An AI champion takes the data your business already has, runs it through AI, and turns what it finds into changes the team can use. Here are four jobs I would give a new champion. Each one uses data you already own and can start this week.
1. Customer service: find what customers love, hate and keep asking for
Have Claude or ChatGPT analyse every inbound and outbound email from the last 12 months, plus call transcripts if your phone system records calls (job notes work if it doesn’t). Gmail and Outlook can both export a folder or label in one go. Ask it for three lists:
- The biggest points of contention. The complaints and pain points that come up again and again.
- What makes customers happiest. The moments people thank you for, so you can do more of them on purpose.
- What customers ask for that you don’t offer. Services you could add without much operational overhead and offer as upsells when people call.
Then turn the findings into scripts and guidelines for your reception staff, so every inbound call becomes a chance to solve the real problem and mention the add-on people already want. Your front desk becomes part of the sales team.
Customer service analysis prompt
Export your emails or call transcripts to a file first, and strip out phone numbers, addresses and payment details before you upload anything.
Analyse them and give me:
1. The 10 most common complaints or points of friction, ranked by how often they appear, with 2 short example quotes for each.
2. The 5 things customers praise most often.
3. Every service or product customers asked for that we don’t currently offer, with how many times each was requested.
Rules:
– Only use what is in the files. Don’t invent numbers or examples.
– If something only appears once or twice, say so.
Then write a one-page phone guide for our reception staff: how to handle the top 3 complaints, and when to mention the top 2 requested add-ons.
2. Marketing: three places your champion should look first
Marketing is where an AI champion pays off fastest, because the data is already sitting in your ad accounts and website tools.
If you want to go deeper on the website side, I wrote up how I run a full SEO audit with Claude, and the five marketing jobs Claude can run for a local business.
I train AI champions on exactly these jobs, using your own data, so they can run them without me. Ask me how it works →
3. Sales: read every field note at once
Claude can connect to business systems through MCP (Model Context Protocol). HubSpot has an official Claude connector, and Salesforce has an official plugin in beta on paid plans, for organisations Salesforce approves. That lets Claude read your records directly instead of you copying and pasting. Your champion can then analyse sales reps’ notes at scale: what are the most common objections, what feedback keeps coming up, and what do customers say just before they buy or walk away?
That turns hundreds of scattered notes into a short list of objections to answer in your quotes, your website and your follow-up calls. If you are a trade business on a job management system rather than a CRM, such as ServiceM8 or simPRO, an export of job and quote notes does the same job.
4. Run it on a loop
Once a job works once, your champion can make it run every day. Claude Cowork on paid plans can run scheduled tasks, so it can look at the same data each morning (tasks run in the cloud, so they keep going when your computer is asleep or the Claude app is closed, unless the task needs files or apps on your computer) and produce a daily report of where the business can be more efficient: new complaints, quotes going cold and, once your ad accounts and Search Console are connected, ads that slipped and pages losing clicks. The champion reads it, decides what matters, and acts on it.
For that to work, Claude needs to know how your business runs. I cover that setup in giving Claude a memory of your business and Claude Cowork for small business.
How much time does an AI champion need?
An AI champion needs about two hours a week, blocked out on their calendar and treated as part of their job. That is my recommendation for small teams, and it matches FIT Automate’s estimate. Lead with AI’s guide, written for larger programs, puts it at 30 to 60 minutes a week, but in a small business the champion is also doing the analysis. A simple weekly rhythm:
- One share-out of something that worked, in the team meeting or group chat.
- One check-in with someone who has stopped using the tools.
- One update to the shared library of prompts that work.
- One report to the owner on what they found and what needs a decision.
- One experiment on a new job or tool.
What goes wrong with AI champions?
AI champions fail when the role is bolted on top of a full job with nobody backing it. These are the common ways it falls over:
- No protected time. If it sits on top of a full workload, it is the first thing dropped in a busy week.
- Burnout. Lead with AI notes many champion programs fade after about six months, especially when the champion becomes the default help desk for “where is the button” questions.
- No data rules. Without clear guidance on approved tools and what customer data can go in, staff fall back on personal accounts. Use a business plan, and check the tool’s data and training settings before any customer data goes near it.
- Measuring the wrong thing. Logins and prompt counts don’t show much. Hours saved, faster replies and more quotes won do.
- No recognition. If the champion’s wins are never mentioned, they stop sharing them.
How do you measure whether your AI champion is working?
Measure what changed in the business. Ask your champion for a short monthly note that answers three questions: what took less time, what we did with that time, and what we found that made or saved money. Lead with AI calls this a reinvestment report. If the answer to the second question is “nothing”, the time saved has leaked away, and that is worth fixing too.
- Your team already uses ChatGPT or Claude with no shared approach
- You have 12 months of emails, ads or sales notes nobody has looked at
- You have a trusted person who could free up two hours a week
- Nobody can free up any time at all
- You have no rules on customer data yet
- The owner isn’t willing to back the role publicly
The honest summary
An AI champion won’t transform your business in a fortnight. What they will do is make sure AI gets used on the problems that cost you money, by someone who understands those problems, every week. Pick the person this week, block two hours in their calendar, and give them the customer service job first. It uses data you already have and usually turns up something worth fixing in the first fortnight.
Straight answers (FAQ)
Sources: Microsoft, Workers worldwide are embracing AI, especially in small and medium-size businesses (2024 Work Trend Index); Lead with AI, AI Champion Programs (2026); FIT Automate, The AI Champion: the quiet role that makes adoption stick in small teams (2026); BizTech, How building an AI champions network can help adoption (2026); Writer, 2025 enterprise AI adoption survey; Fortune on MIT NANDA, The GenAI Divide (2025).
Put an AI champion in your business.
I help trade and service businesses choose the right person, set up the tools and data rules, and train them hands-on to run the jobs in this guide on your own data. When we’re done, the skills stay in your business.
Happy to do it yourself? Everything above is the real playbook. Use it. Not sure where your marketing is leaking first? Take the free scorecard →
Get in touch
Prefer to talk it through? Call or text, email, or send a message.