By Kim Leary
The short version
- Most teams start with the tool and look for a use. It works far better the other way round.
- The AI that sticks is built around a job someone already does. The demos rarely are.
- Starting with the tool leaves you with an objective, a budget and not much to show for either.
- The common approaches spread the effort too thin to change anything.
- Where your team’s week goes tells you where to start.
- Pick the two or three jobs that cost your team the most, and set AI up properly for those, on your brand and your data.
It’s in your objectives this year: make the team more efficient with AI. There’s even a budget line for it. You use it every day already, for emails, summaries and first drafts. And yet you still haven’t found the thing that changes how the team works.
Meanwhile you’ve been handed a new market with no extra people, the tools change every month, and most days the job is keeping your head above water.
Why it happens
Most teams start with the tool and look for a use. It works far better the other way round.
The question isn’t “what can AI do?” It’s “what’s taking my team the most time, and could it be done better?” Asked that way, the answers are usually unglamorous: the monthly report, the content that gets made once and never reused, the admin between systems, the brochure that takes a designer a week.
Those are the places AI earns its keep. Not because it’s faster, but because it makes room for work that used to get cut.
What it looks like when it works
The AI that sticks is built around a job someone already does. The demos rarely are.
One growth organisation we spoke to paid an AI consultancy for six months. They got a lot of demos, a couple of builds that never really worked, and one thing that was genuinely useful: an internal agent that works from their own data.
A solo marketer we spoke to connected an AI assistant to her project management tool, to plan and sequence the team’s tasks. It isn’t the kind of thing anyone puts in a demo. It was a real job, and it worked well.
Both point the same way. Start with the job, and the right tool usually follows.
What it’s costing you
Starting with the tool leaves you with an objective, a budget and not much to show for either.
What doesn’t fix it
The common approaches spread the effort too thin to change anything.
“Give everyone a licence and see what happens.” You get lots of small, private uses and nothing you can show.
“Wait for the killer app.” It’s unlikely to be one tool. It’s more likely to be your process, set up properly.
“Pick the best tool.” The best tool changes every month. The setup that works for your team doesn’t.
“Use it to make more content.” More of the same, in a market that’s already full of it.
Which job to start with
Where your team’s week goes tells you where to start.
It’s probably production if:
- Sales, events and leadership all want decks and one-pagers, and they all come through you.
- People already use AI, and what comes back is nearly right and slightly different every time.
- Your brand is getting quietly diluted one deck at a time.
It’s probably reporting if:
- Every report takes a day to pull together, and nobody agrees on the numbers.
- The data is spread across tools that don’t talk to each other.
It’s probably direction if:
- Nobody has agreed what AI is for across the business, only in marketing.
- There’s a marketing person, but nobody setting the direction above them.
Whichever list sounds most like your week is the job to start with. Each one has its own fix below.
What fixes it
Pick the two or three jobs that cost your team the most, and set AI up properly for those, on your brand and your data.
AI tools and training starts with your design system and messaging, then sets up the tools, prompts and guardrails around the work your team actually does, with half a day of training on your own work. A one-pager or brochure that used to take a week comes out on brand in an afternoon. From £2,800 + VAT.
If the job eating the most time is reporting, connected data and reporting joins the tools so the report builds itself. And if you need to agree what AI is for across the business, not just in marketing, that’s growth consultancy.
Keep reading
- Why AI-generated graphics all look the same
- Most of your content gets used once
- What good AI advice looks like
Questions about how to use AI in a marketing team
Start with the jobs that take the most time, like reporting, reusing content and producing on-brand materials. Set AI up for those, on your brand and data, rather than giving everyone a tool and hoping.
It changes every few months. What lasts is the setup: your brand, your messaging and your templates built into whichever tool you use.
Pick a specific job, measure how long it takes before and after, and show what the team did with the time.
A good one names a job and what better looks like, not a tool. For example, a monthly report that answers the board’s follow-up questions, or sales materials the team can make on brand themselves.
If you want the output to look like you, yes. Without one, the tools fill the gaps with their own defaults. AI tools and training starts there, and if you haven’t got one, building it is the first part of the job.
On the setup before the subscriptions: the brand, messaging and templates the tools work from, and training on your own work. Before you spend it, what good AI advice looks like covers how to buy help without being sold a tool.
Free resource
Build the business case. Put a number on the job you want AI to do better, before you spend the budget line.
The one thing to do next
AI tools and training
Your design system and messaging built into the tools, prompts and guardrails, then half a day of training on your own work. From £2,800 + VAT.
