Before the three months
Discover
We map how work moves through the agency, assess your team's skills and find where the hours go. Then we set a target for each workflow and get your client-data rules signed off.
AI training for
marketing agencies
marketing agencies
Make AI part of how your agency works.
We train your whole team to use AI on the client work they already do. A programme designed for your agency, run inside it for three months.
Book a call ScrollYour team is flat out on client work, so learning AI keeps getting pushed back. We teach it inside that work, until it's simply how things are done.
Our founder has worked with
Sound familiar?Marketing, creative and digital agencies that already pay for AI and want everyone using it.
- 01
The licences are paid for.
The invoices land every month, and day to day the team works the way it did two years ago.
- 02
Two or three people are flying.
They use AI brilliantly. What they know lives in their heads and their browser tabs, so it never spreads.
- 03
Training didn't stick.
The team went to a workshop and came back keen. A week later the work had piled up and the old habits were back.
Built around your agency.
Every programme is designed for your agency. It starts with discovery, runs for three months on live client work, and finishes by measuring what stuck.
The three months
Build
Your people rebuild their own workflows with us, on live client work. Each one is tested on past jobs, gets an owner and gets written down, and we coach until it holds up in a busy week.
After the three months
Measure
We time each workflow and count its rounds of changes, before and after. Once we've stepped back, we check adoption is holding and fix anything that hasn't stuck.
Your numbers
What does routine work cost your agency?
Set three numbers for your agency. You'll see how much of each month goes on routine work, and what it costs you.
Routine work is the work that follows the same pattern every time, the kind AI can take a first pass at.
- Status reports
- Meeting notes and follow-ups
- First drafts of briefs and copy
- Client reporting
- Resizing and versioning assets
- Proposals
- Timesheets and admin
Everyone on the payroll, including leadership.
Think of a normal week and add up the tasks above.
Average salary plus overheads, per hour worked.
Routine work, every month1,040
hours a month. That's about 6.4 full-time people.
What those hours cost, every month£46,800
How it's worked out: people × hours × 52 weeks ÷ 12 months. A full-time month is 162.5 hours, based on a 37.5-hour week.
The founder
We've done this before.
Ideal and Digital was founded by Gavin Pieterse, who has trained people since 2017. At Multiverse he ran programmes for Google, Sky, Meta and GroupM and led a team of over 40. Today he teaches AI to teams through workshops and programmes.
Ideal and Digital brings that experience into agencies. We work inside your team until AI is part of how it works, and we measure the change.
- 2017
- Training people since
- 1,500+
- People through his programmes
- 40+
- Team he led at Multiverse
Insights
All insights
AI Implementation
Claude Dashboards and Motion: what agencies should test
Claude Dashboards and Claude Motion launched in beta on 8 October. What each one does, which plans get them, and the first jobs an agency team should test.
Read the article
AI Strategy
Is it safe to put client data into AI? A plan for agencies
Before you put client data into AI, check your plan, your client contracts and the UK GDPR basics, then adopt a simple one-page policy for your agency team.
Read the article
Team Adoption
AI adoption in agencies: who moves first, where it stalls
AI adoption in agencies tends to follow one path: who moves first, where it stalls, what the quiet majority does and what gets the whole team using it.
Read the article
Questions we get asked.
Who is this for? +
Marketing, creative and digital agencies that already pay for AI tools and want the whole team using them on real client work. It's built for agencies of about 20 to 100 people. If yours is bigger, we'll plan for that in the proposal.
Which AI tools do you train on? +
The ones your agency already pays for, whether that's ChatGPT, Microsoft Copilot, Gemini or Claude.
How long does it run? +
Three months of training inside your agency. Discovery comes before it and a check on adoption comes after, all shaped around your teams and your calendar.
What will we see, and when? +
The first three weeks are for learning the tools on your own work. The first time savings should show by weeks four to six, and the new ways settle in by weeks eight to ten. If a workflow hasn't moved by the halfway point, we swap it for another one, so the three months don't stretch.
How is this different from an AI workshop? +
It runs for three months inside your agency, on live client work, so your team practises on real jobs every week. We coach until the new habits hold up in a busy week, and we time each workflow before and after so you can see what changed.
How do you measure the change? +
Before we start, we time each workflow, count how many rounds of changes it takes and agree a target with you. Each new way is tested on past jobs and runs alongside the old one until it's proven. At the end we time it and count the changes again, check who's using the new way and ask the team whether they'd go back to the old one. You see the difference in hours, and what it means for your margin or capacity.
What's left when you step back? +
You keep each workflow written down, with an owner who can run it and teach it, and the instructions behind it in a shared library your team can update. You also keep the signed client-data policy and a list of what to tackle next, built from ideas your team had along the way.
How much of my team's time does it take? +
Your team learns on the work it already has, so most of the time goes into client work. Each workflow owner also needs a little protected time every week. We agree the commitment with you, team by team, before we start.
Is it on site or remote? +
We agree the mix of on-site and remote work with you, based on how your teams work and where they are.
Is it safe to put client data into AI tools? +
It can be, once the rules are clear. Each programme includes a one-page client-data policy, signed off by your managing director before the team starts, so everyone knows what can go into which tool and what a person checks before a client sees it. Our guide to client data and AI covers the basics.
What happens on the first call? +
We talk about where your team is with AI, what's getting in the way and whether a programme would help. If it would, a written proposal and price follow the call. There's no obligation.
What does it cost? +
Every programme is shaped to the agency, so we price it after the first conversation.
Let's talk.
A short call about where your team is with AI and what's in the way. You'll come away with a clear next step either way.