What a marketing leader told us about AI and the numbers that matter

A firm should judge AI by the numbers it already runs on: its recovery rate, fee income and gross profit. A figure that only looks good in a report tells the partners nothing about whether the money came back, and Charlie Barrett made the same point about marketing when we spoke to him last month.

Charlie runs cbc., a consultancy that gives industrial and technical businesses senior marketing leadership without a full-time hire. He has spent years leading marketing and ecommerce inside manufacturers, and his work keeps coming back to one question: can each piece of marketing be matched to a result? His own site promises no fluff and no vanity metrics.

A big percentage can hide a small number

Charlie often finds the same thing when he starts with a new client: "the agency sold them the dream, they've given them some really nice, fluffy reports." One report told a client, "this month we've grown your AI assistant traffic by 300%." The numbers behind it were small. "It had gone from 6 to 18," he said, at a business that "needed 187 sessions before they got a qualified lead."

AI often gets reported to partners the same way, as hours saved or prompts sent, and neither figure says whether the firm earned more. The recovery rate does. Every job has a time budget, and the fee is set from it. When a job runs over, the extra time is written off, so the firm bills less of the time it recorded. When AI brings jobs in inside their budget, less is written off, and the recovery rate, fee income and gross profit rise together, while the same team takes on more clients without adding to its headcount. So every solution on a roadmap we write is costed against those numbers before anything is spent, and checked against them once it is live.

The return starts with knowing where the money goes

Costing a solution that way depends on knowing where the firm makes and loses money today. Charlie's own focus is "where you spend money, where you get customers, what those customers are worth", and with that in place, he said, "you can articulate quite quickly what is wasting money, what's making money."

When we spoke he was reworking a client's Google Ads, doing much of it in Claude after giving Claude the context of the business, and he was careful about the balance: "So it's all supported by AI, not driven by it."

An AI that knows nothing about a firm produces generic work, so we check the firm's own context before any solution goes on the roadmap. Once a solution is live, a named owner still makes the judgement and checks what it produces before it reaches a client, so the standard of the work holds or rises.

The hours worth most hide in everyday tasks

Checking that context is also where the hours worth most turn up, often in an ordinary task the team repeats every week. That same client wants to move to Xero, with a quoting system in HubSpot connected to it, because at the moment "it takes them an hour to do a quote because they do it in Word." Charlie's view was that "if we can get the ecosystem right, actually that would be massively beneficial for the business", as long as it doesn't take anything away from what they already have, since it is "quite a lean business anyway."

Much of the time we find in firms turns up like this, an hour at a time. Inside a client job it is the hour that pushes the job past its time budget and gets written off. Outside one it is admin eating into chargeable time. Either way, freeing it gives the same team more capacity.

Getting the ecosystem right comes first: a tech stack that works with AI, the steps written down, the data organised and the systems connected. Then the solution runs where its data allows. A quote carries client names and prices, so it belongs on the firm's own server, where the provider never trains on what it sends, or on a machine in the office. Only work with nothing sensitive in it goes to the provider's own cloud.

To know which everyday tasks hold the hours worth most in your firm, and how you would know AI had paid, take the quiz below. It takes a few minutes.

Most firms need the foundations before the agents

How far a firm can go with AI depends on where it starts. "I think there's this massive gap, isn't there?" Charlie said. On one side are people already running agents, and on the other are people who have no idea at all. He told us about a friend whose business turns over millions with all of its contacts in a spreadsheet, and what he said to him: you can "implement AI that can be not scary implementation, just bog standard implementation that can help support."

Most of the firms we speak to sit nearer that second group, and for them the foundations come before any agent, with a plan that says which solution comes first and how much time it stops them writing off. For the firms further along, the risk is a solution drifting when the model underneath it changes, so every solution we implement is checked and tweaked as the models change, and the quality of the work doesn't drop unnoticed. Because the firm owns its knowledge and rents the AI, it can move to a better model when one comes along.

What firm leaders should do now

Taken together, Charlie's points set out the order a firm needs to follow, whoever does the work.

Start with the firm's own work, broken down into the individual tasks and processes that create value for its clients: who does each one, how long it takes, what data it uses and who signs it off.

Before any solution, check the layers those tasks sit on: the business and each department, the tech stack, the quality, completeness and security of the data, compliance and governance, and the people who will work alongside the AI. Where a layer isn't ready, fixing it comes first.

Then set out a roadmap of solutions, each fully costed with the hours it frees up, what it costs to implement, its first-year return and what it does to the recovery rate on the jobs it touches.

Implement the solutions one by one, in the roadmap's order, each running where its data allows, with its risks designed out, covered by the AI policy and owned by a named person who checks what it produces. Once they are live, keep managing them as the models change.

Done in that order, the firm gets more capacity from the same team, hits its chargeable time targets and lifts its recovery rate and fee income, in a way that is compliant and keeps client data secure.

If you want us to do it for you, book a free call and we'll find you at least $50,000 a year in new business that AI can unlock. It's compliant, data-secure and has governance built in.

Know exactly where AI would pay in your firm, in hours you can count

Our quiz asks where your team's week goes, how often jobs run over their budgeted hours, where client data ends up and how you would know whether AI paid. Your report shows where AI would free up your team's time first, the capacity it could give your firm in hours, and a 30-day plan for what to do.

We'll find you $50,000 a year in new business AI can unlock in 72 hours, or you don't pay.