Using AI in Marketing Without Producing Mediocre Work Faster
Output has risen in every marketing team that adopted the tools. In most of them the work got faster and no better, because the constraint was never production.
Theresa Ronnie · 23/10/2026 · 5 minute read
AI adds capability to an in-house marketing team in four places: research synthesis, variant testing, first drafts and measurement. It does not supply the strategic question, taste, or the decision in the approval room. Because AI compresses production, the constraint moves upstream to the brief and the judgement applied to it. Teams that automate output without fixing the brief produce more mediocre work faster. Fix the brief first, then automate.
For most of the history of marketing, production was the bottleneck. Ideas were cheap and making them was expensive, so teams were organised around making: studios, agencies, timelines, approval gates. AI has compressed that. A first draft of anything now takes minutes. Fifty variants of a headline take less time than one used to. Research that took a fortnight is summarised before lunch.
When a bottleneck is removed, the constraint does not disappear. It moves. In marketing it has moved to the brief, to the judgement applied to what comes back, and to the room where the decision is made. Teams that have not noticed this are using AI to produce more of what they were already producing. The result is predictable: more mediocre work, delivered faster, reaching the market before anyone has judged it.
Why the brief matters more now, not less
A weak brief used to be expensive to execute, which gave people time to notice it was weak. Now a brief that asks for "a campaign for the launch" returns a campaign for the launch in an afternoon, complete and plausible, with nothing in it that a business could measure. The brief was never challenged because nothing in the process required it to be.
The result is that the quality of the brief is now the quality ceiling of the work. A brief that names the commercial number, the date, the audience by behaviour and the single question to answer will get usable work from AI and better work from the agency. The tools are good at answering questions and poor at inventing them. A brief without a question gets volume.
Where AI adds capability inside a team
Research synthesis. Customer interviews, sales call notes, reviews, category reports and competitor messaging, read and summarised in hours rather than weeks. This is the part of strategy work that used to be skipped because there was no time. There is time now, and a team that uses it writes better briefs because the brief is finally informed by something.
Variant testing. Twenty versions of a message tested against a live audience before the campaign commits, with the losers discarded before anyone has fallen in love with them. This replaces opinion with evidence at the point where opinion used to be most expensive.
First drafts. The blank page is gone. A team that treats the first draft as a draft, something to be argued with and improved, gets to the better idea faster. A team that treats it as the deliverable ships the median.
Measurement. Connecting marketing activity to commercial outcome used to need an analyst the team did not have. The analysis is now within reach of the marketing manager, which means the gap between what marketing did and what the business got can finally be closed by the people who need to see it.
Where it does not
The strategic question. Which segment, which proposition, which problem is worth solving this quarter. The tools answer the question they are given. Give them the wrong one and the answer is wrong at scale.
Taste. The tools produce the statistically likely version, which is by definition the one the audience has seen before. Choosing the distinctive idea over the probable one is a human choice, and it needs a person willing to argue for it.
The decision in the room. No tool approves work against the brief rather than against seniority, and if the room approves on taste, AI gives it more options to approve badly. How the argument and the approval are run is covered in Better Creative Output Is Decided Before the Studio Starts Work.
How creativity goes up
In the teams we have seen improve with AI, the gain did not come from producing more. They used the production hours they recovered on the parts of the work the tools cannot do: a sharper brief, a longer argument about the idea, a harder question in the approval meeting. Creativity rose because judgement finally had time. The tools did the drafts. The people did the thinking, and automated only once there was something worth multiplying.
In the Constraint to Capability Growth System, the question asked of AI is not which tools the team should adopt but where the constraint has moved to since they were adopted.
How should an in-house marketing team use AI?
An in-house team should use AI where it adds capability: synthesising research for the brief, testing message variants before committing, producing first drafts to be argued with, and connecting marketing activity to commercial outcomes. The strategic question, taste and the approval decision remain the work of the team.
Why has AI not improved our marketing results?
AI has not improved results because production was not the constraint. The constraint moves upstream to the brief, judgement and the approval room, so a brief that asks for an asset gets more assets faster. The number stays flat because nothing in the system asked it to move.
Does AI make marketing more creative or less?
AI makes marketing more or less creative depending on where the recovered hours go. Creativity rises when the team spends the time saved on a sharper brief, a longer argument for the distinctive idea and a harder approval question. It falls when the time saved goes into more volume.
What should a marketing team fix before adopting AI tools?
Before adopting the tools a team should have a brief with a number and a question, an approval that is run against that brief, and one measured line from activity to outcome. Without those, the tools multiply whatever the team was already producing.
Will AI replace the marketing team or the agency?
AI replaces production hours, not judgement. The strategic question, the choice of the distinctive idea over the safe one and the decision in the room remain human work. A team that treats the first draft as the deliverable will find its output indistinguishable from everyone else using the same tools.
Bring us the number that will not move.
A thirty-minute Creative Growth Diagnostic with Theresa Ronnie locates the constraint and puts a rupee figure on it. If the Growth System does not fit, we say so on the call.
- A Marketing Team Upgrades in 90 Days by Changing How It Works, Not Who Is in ItHiring and reorganising are the visible fixes and the slow ones. What changes inside a team in 90 days when the method is practised on live work instead of taught in a workshop.
- A Busy Marketing Team With Flat Growth Is Rarely the ProblemA marketing team that is busy but not moving the number is usually constrained in one of five places. How to tell which, before you change the team or the budget.
- Better Creative Output Is Decided Before the Studio Starts WorkCreative quality is decided before the work is made: in the brief, in the argument for the idea, and in the room where it is approved. Fix those three and the output follows.