AI Consulting
AI built into how your team already works, so you ship faster and cheaper without losing the craft.
For in-house marketing and creative teams. Run remotely, inside your own tools.
Everyone has the tools. Almost nobody has the workflow
Your team has licences. Someone ran a pilot. There is a shared prompt doc nobody opens. Output rose for a fortnight and then settled back to where it was, because none of it changed how the work actually moves.
The failure is rarely the model. It is that AI got bolted onto a process built around the old constraints, so it saves twenty minutes on a task that was never the bottleneck in the first place.
And the output gives itself away. Generic phrasing, strange detail, a house style quietly dissolving. The team notices, loses confidence, and goes back to doing it by hand.
A tool no workflow depends on is a subscription, not a capability.
What we do
We are not here to demo software. We are here to change where your work gets stuck.
Map how the work actually moves
Not the process diagram. The real one: who waits on whom, where revisions pile up, and which single step sets your ceiling.
Fit the stack to the job
We choose tools against your actual deliverables and drop the rest. Fewer tools used properly beats a stack nobody has time to learn.
Build reusable systems
Prompts, references, review checklists and asset templates that produce a consistent result whoever on the team is running them.
Train and hand over
Your team runs it after we leave. We document the system and work through live briefs with the people who will own it.
How it works
Five steps over four to six weeks. The pilot is a real deliverable, not a demo.
Map
Sessions with the people doing the work, tracing one real deliverable from brief to publish.
Find the ceiling
We identify the step that actually caps throughput. It is usually not the one people expect.
Pilot
We rebuild one real deliverable with AI in the loop, and compare it against your current baseline.
Systemise
The pilot becomes documented prompts, references and quality gates your team can repeat.
Hand over
Training on live briefs, then a review after a month of your team running it alone.
What you get
Documentation your team keeps, not a slide deck about the future of work.
- A written map of your current creative workflow and its bottlenecks
- A recommended tool stack, scoped to what you actually produce
- A documented prompt and reference system for your brand
- Quality gates that catch AI output before it reaches a feed or a client
- One deliverable rebuilt end to end as a working reference
- Live training sessions with the team who will own the workflow
- A written playbook that stays with you
- A follow-up review after a month of independent use
We run this stack ourselves, every day. See what it produces.
Proof
We teach the workflow we run on client production.
Per finished video asset, using this workflow
The turnaround we hold on client production, against the three to six weeks a traditional UGC batch takes. The consulting work is teaching your team the same system.
Read the case study Post-PeakLocalisation without a second shoot
Four videos localised for a new market in days, where casting voice actors and re-editing lip sync would have added two to three weeks.
Read the case study FAM HOMEA studio line removed from the budget
Amazon.de product video produced without a shoot, keeping the shelf presence and cutting the production cost.
See the workWe publish our own production numbers and nothing else. We do not quote client cost savings we cannot source, and the gain you see depends entirely on where your bottleneck actually sits.
AI consulting — frequently asked questions
Usually less than you expect. Most teams we work with already hold more licences than they use. We audit what you have first, and only recommend something new when a specific deliverable needs a capability you genuinely lack.
Four to six weeks for the core work, plus a review a month after handover. Longer engagements exist where a team wants us to rebuild several workflows in sequence, but we would rather prove one before starting the next.
That is the risk we spend the most time on. It is why the deliverable includes quality gates and reference systems rather than just prompts. Anyone can generate an asset. The judgment about which one is good enough to publish is the part that has to stay human, and that part is what we train.
The people doing the work, not just the people managing it. Realistically that means a marketing lead plus two or three of the designers, editors or writers whose day the workflow will change. Roughly two hours a week during the engagement.
Teach. This service ends with your team independent. If you would rather we produced the work instead, that is Creative Production.
Make AI part of how you work
One call to find the step that is actually capping your output.
Book a workflow call