A roadmap
The most valuable opportunities across all three horizons.
Brand & MarCom
Make AI pay off in marketing and sales.
We work with marketing, communications and sales teams to turn their strongest AI opportunities into prototypes that hold up in the real workflow.
The situation
Gen AI made the work faster.It did not make it land better.
Drafting got cheap. Reach, response and pipeline did not move with it.
More tools. More systems added to the work, with no agreement on where they sit in the process. Each person has their own way in, which means the team has none.
More experiments. More pilots without ownership. Something works for one person, but the team's way of working does not change.
More output. More content, not always more value. Volume rose because drafting got cheap, and review absorbed the time that drafting saved.
And the hours that did come back went nowhere, because nobody decided in advance what they were for.
The question is which work should change, and who owns the value it frees.
The method
A 12-week cycle turns potential into working proof
We embed in the Brand & MarCom team, find where AI can create the most value, and turn the strongest opportunities into 4–6 working prototypes.
Where can AI create the most value here?
We map the work the team actually repeats: briefs, drafts, reviews, reporting, hand-offs to sales. Then we rank it by value, friction and feasibility. The shortlist is built from your work, not from a category benchmark.
Which opportunities become prototypes?
Four to six, chosen with the team and spread across the three horizons: better work today, a redesigned process, and something new the brand could sell or offer. Small enough to build, real enough to judge.
What proves that it works?
Each prototype is built on real people, real data and real workflows, and measured on quality, time, capacity, customer value or job satisfaction. Evidence that shows up in the work itself.
Can the team keep building after we leave?
We run the first cycle together and hand over what works with a named owner. The measure we hold ourselves to is how much your team ships without us.
What you get
We leave a capability, not a deck.
People can use and own it, the business value is visible, and the technology is ready for the next step.
4–6 prototypes
Built with real people, data and workflows.
Proof of value
Evidence across quality, time, capacity, customer value or job satisfaction.
A backlog
The next opportunities, ready for another cycle.
A clear decision
Stop, improve, or scale.
Evidence
The productivity is real. The return is not automatic.
Faster does not automatically mean more valuable. So we build, test and measure. Then we decide. That is why this is a cycle: to turn AI potential into better decisions.
McKinsey, The state of AI in 2026, 25 August 2026.
What we could build together
Every cycle picks 4–6 prototypes.
These are examples. Your prototypes will come from your team's own work and priorities.
Create content fasterContent repurposing engine
One asset becomes a series, on brand, without the work.
Improve brand consistencyBrand voice writing assistant
A writing tool that applies your tone-of-voice rules before review.
Turn customer data into insightCustomer insight synthesiser
Every interview, survey and ticket, readable at a glance.
Equip sales more effectivelyProposal and pitch assistant
Sharp proposals built on company, brand and competitor information.
Understand performanceContent performance analyst
What worked, why, and what to do next week.
Improve AI visibilityLLM brand monitoring tool
Watch how AI models perceive your brand and your competition.
See 26 more prototype ideas
- Campaign concept generator
- AI video production workflow
- Multilingual content engine
- Competitor intelligence agent
- Landing page generator
- Thought leadership engine
- Marketing dashboard
- Slack signal digest
- Customer case writing assistant
- Event campaign workflow
- Interview and transcript intelligence
- Cross-sell content agent
- Email campaign and newsletter assistant
- Social content studio
- Rapid visualisation studio
- Creative testing simulator
- Interactive brand utility
- Product recommender
- Customer journey optimiser
- Brand asset search engine
- Knowledge-to-content engine
- Vacancy content generator
- Always-on editorial planner
- AI-powered market researcher
- Sales story builder
- In-house video studio
AI visibility
Will AI recommend your brand?Or a competitor?
For twenty years, search meant competing for a place in a list of links. An AI answer compresses that list into a short recommendation. Your brand is in it, absent from it, or described wrongly in it.
When a buyer asks an AI model who to hire, what to buy or which supplier to trust, the answer is assembled from what it can find and interpret about your brand, including sources that are stale or contradict each other. Most brands have never checked what that answer says.
What the scan and the build cover
The scan: where you stand- The prompts buyers in your category actually use
- Whether you appear, and how you are described when you do
- Which competitor is named instead, and why that one is easier to resolve
- What the models have wrong: stale facts, old positioning, products you no longer sell
- Which sources they are drawing from: your site is only one voice
- Machine-readable identity: structured data, one consistent description everywhere, verified profiles
- Answers to the questions buyers type, written so they hold up when quoted out of context
- Presence on the third-party sources the models read
- The same prompt set re-run monthly against the baseline, so movement is visible
What we will not promise: that a model will cite you. Nobody can, and the mechanics change every quarter. What we can do is make your brand easier for models to identify and understand, keep your facts current, answer the questions your buyers ask, and measure the difference over time.
MarCom sits across all three horizons.
Better work today is adoption. The process around it is redesign. And a brand that can offer something it could not offer before is a new business model.
Questions we hear
Asked in every first conversation
The five that come up before anyone talks about scope.
We already use AI for content. What is different here?
Drafting got cheap, and that is where most teams stop. Reach, response and pipeline did not move with it. We do not add another tool to write faster: we find where AI changes what the work produces, build it, and measure whether it landed.
Twelve weeks is a long commitment before we know this works.
It is not the only way in. A scan finds and ranks the opportunities before anyone commits a cycle. A challenge puts people and AI on one real brief. A sprint builds a single prototype. Each ends in a decision, not a proposal for the next step.
Who builds these, your people or ours?
Yours, with us in the room. We embed in the team for the cycle and build on real people, data and workflows. What stays behind is a capability your team can use and own.
How do you decide what to build?
We map the work the team actually repeats—briefs, drafts, reviews, reporting, hand-offs to sales—and rank it by value, friction and feasibility. The shortlist comes from your work. Four to six become prototypes, chosen with the team.
Can you get AI models to recommend us?
No, and nobody can: the mechanics change every quarter. What we can do is make your brand easier for models to identify and understand, keep your facts current, answer the questions your buyers ask, and re-run the same prompt set monthly so movement is visible.
Who you work with
Mathijs Boonstra
Brand & MarCom
He brings ideas to life with marketing, communications and sales teams. He combines strategy, creation and rapid prototyping to develop workflows, tools and content that would not have been possible before.
Over the last 24 months, he has worked with 16 teams on 36 prototypes.
Mathijs worked with the Marketing & Communications team of a leading accountancy firm to explore AI opportunities. AI came alive when we started making choices. We prioritised the strongest opportunities and turned five into active pilots and demos, moving AI from possibility to daily practice.
Ways to start
Four ways to start.
An idea. A question. A workflow. A challenge. Twelve weeks is not the only way in, and the smallest step still ends in a decision.
AI Opportunity Scan
First stepFind and prioritise the opportunities with the greatest potential, before committing a cycle to any of them.
On site + remoteInterviews + your own material
What you get
- The work your team repeats, mapped and ranked
- A shortlist scored on value, friction and feasibility
- Which of them are worth a prototype and which are not
- A go or no-go on committing a cycle
Teams that suspect there is value here but cannot yet say where.
Fixed feeStated in the first conversation, for a defined deliverable.
Talk about the scan (opens in a new tab)
Creative AI Challenge
Put people and AI to work on one real brand or customer challenge, and see what the pairing produces.
On siteYour team + Mathijs
What you get
- One real challenge worked on with AI in the room
- What the pairing produced, and where it fell short
- A view on which work is worth automating and which is not
- The team’s own read on what changes next
Teams that learn by doing rather than by deciding first.
Fixed feeStated in the first conversation, for a defined deliverable.
Talk about a challenge (opens in a new tab)
Prototype Sprint
2 to 4 weeksBuild one promising opportunity and test it in practice, with the people who would use it.
Hybrid2 to 4 weeks5 to 15 participants
What you get
- One prototype running on real work
- Your people in the build from day one
- Evidence on quality, time or capacity
- A decision: stop, improve, or scale
Teams with one idea they already believe in and want tested.
Fixed feeStated in the first conversation, for a defined deliverable.
Talk about a sprint (opens in a new tab)
The 12-week cycle
12 weeks · the full programmeWe embed in the team, find where AI can create the most value, and turn the strongest opportunities into 4–6 working prototypes.
On site + remote12 weeksThe Brand & MarCom team
What you get
- A roadmap across all three horizons
- 4–6 prototypes built with real people, data and workflows
- Proof of value, and a backlog ready for another cycle
- A clear decision: stop, improve, or scale
Teams ready to move from experiments to a capability that stays.
Fixed feeStated in the first conversation, for a defined deliverable.
Talk through a cycle (opens in a new tab)
LLM Brand Scan — where your brand appears, disappears or is misunderstood in the models your buyers ask. See AI visibility above.
Start with one workflow. End with a working prototype.
Tell us which piece of work your team repeats most, and what you would do with the capacity if it came back. Some of what lands on this page turns out to be a process problem, not a marketing one, and that is worth finding out in an hour.
Talk through a cycle (opens in a new tab)
30 minutes with a founder. No deck.
Not ready to talk? Take the 5-minute AI Horizon Scan →