Home · AI Consulting

Forward deployed AI engineers

We automate your business with AI.

First we map how your business actually runs. Then we build the AI version: agents that do the repetitive work, with your people in charge. You see what to automate first, and what it saves, before you spend on a build. Start with a free AI opportunity assessment: see what to automate first, and what it saves, before you spend anything.

Approved API developer: Google · Meta · TikTok

The problem

Your best people are doing work a machine should do.

Somewhere in your company right now, a smart person is copying data between systems. Another is rebuilding the same report they built last week. Another is triaging an inbox, chasing an invoice, or re-keying a CV into a tracker. You cannot hire your way out: every new hire adds the same manual load that created the problem, and the IT backlog is already six months deep.

A junior ops hire who spends half their week on manual workflow~£17k/yr wasted
One hour of a senior person's day on reporting and reconciliation~£11k/yr wasted
Leads answered in hours instead of secondsdeals lost, unmeasured

Every new hire adds the same manual load that created the problem.

How it works

We map your business. Then we build the AI version.

Forward deployed engineering is how Palantir, OpenAI and Anthropic get AI into production: put the engineer inside the business, not in a slide deck. With Maaten it works in three moves:

01 · MAP

We map how the work really happens

A named Maaten engineer sits with the people who do the work and maps your processes: hours, costs, error rates, systems. You get the Automation Map: every workflow ranked by what automating it returns.

02 · BUILD

We build the AI version

Your workflows get rebuilt as agents on the Maaten AMS, shipped into your live environment, not a sandbox. The first small one is built free during the audit, so you see it working before you commit to anything bigger.

03 · RUN

We run it, you stay in charge

Anything that spends money, contacts a customer or changes a system of record waits for your approval. Every month you get a proof report: hours returned, errors down, one number you can defend.

What to automate

The best work to automate, by department.

The highest returns come from work that is high-volume, repetitive and runs on structured data: fewer errors, faster cycles, and your people freed for the work that needs judgment. This is where we look first:

Tell us which department matters most to you →

Why it holds up

Built on our factory, not on duct tape.

Most automation shops wire together no-code tools and hope. Maaten runs the AMS, our Agent Management System: a production agent platform built by the same team whose media platform holds approved API access with Google, Meta and TikTok. Your agents get monitoring, approval gates, audit trails and measurement as standard.

The audit's first job is making your business readable to agents: your data, rules, prices and permissions gathered into one governed context layer. That layer stays yours, every week of signal makes the next automation smarter, and if you ever leave, the documentation and access go with you.

Your teamsets direction, approves what matters
Specialist agentsdo the repetitive work, 24/7
Maaten AMSbuilds, runs, monitors every agent
Governed data layermeasures everything · yours to keep
Your systemsCRM, inbox, docs, ads, finance

Proof

One number you can defend, every month.

One person now runs international marketing that used to take a team.

An international streaming platform · production deployment on Maaten

80%reduction in data costs
500%faster campaign setup
24/7agents on every account
95%

of internal AI pilots show no measurable P&L impact. The workflow was never baselined, so nobody can prove what changed.

MIT, State of AI in Business, 2025
2x

Specialist partners succeed roughly twice as often as internal AI builds.

MIT, 2025
4-5x

ROI reported on successfully automated workflows.

McKinsey, 2026

Case studies and what a proof report looks like →

Pricing

What it costs.

Most firms in this category quote only after a discovery call. The prices are below.

Step 1 · AI Opportunity Assessment
Free about a week

We map your five to ten most manual workflows and hand you the Automation Map: every workflow ranked by annual saving. Includes your first working automation, built free during the audit. Fee credited in full against your first build.

Step 2 · Pilot build
£15-35k scoped from the audit

Your highest-payback workflow rebuilt as agents and live in production within 30 days, with approval gates you control, measured against the audit baseline.

Your options

The alternatives, compared.

Hire another personDIY no-code / RPABig consultancyMaaten
Time to live3+ months to hire and rampWeeks of evenings, then it breaksMonths of strategy firstFirst workflow in 30 days
Upfront costRecruiter fees + salary"Free", plus everyone's time£200k+ engagementsFree assessment; £15,000-£35,000 build
Ongoing£35-50k/yr, scales one head at a timeLicense creep + maintenance nobody ownsChange requests£2-8k/month, cancel after month 3
When systems changeThey cope, manuallyIt silently breaksNew statement of workWe fix it; that's the retainer
Proof it workedGut feelNobody measuresA closing deckMonthly proof report, baselined

Who you're working with

Built by people who did the job.

Our founders individually ran $500m+ of ad spend for brands including Nestlé, Unilever, Diageo and J&J over 20+ years, before Maaten existed. Together they built a governed data platform approved by Google, Meta and TikTok, and the agent platform that runs on it.

JT
Jay Thomson

Founder. 20 years in digital, $500m+ managed for Nestlé, Unilever, Diageo, J&J.

SvW
Stephan van Wyk

Data OS architect. Secured Meta, Google and TikTok API approval.

RF
Robin Fishley

Agent OS. Ex-Global Director, Search & Analytics, Saatchi & Saatchi.

Common questions

FAQ

What does an engagement cost?

The assessment is free. Pilot builds: £15,000-£35,000, scoped from the assessment. Retainers: £2,000-£8,000 a month depending on tier. Enterprise engagements are priced phase by phase: discovery, strategy, sign-off, then MVP build per business unit. The ranges are real; the exact numbers come out of the audit.

How fast is "live"?

The audit takes 2-3 weeks. Your first production workflow ships within 30 days of starting the build.

Why not just use Zapier or ChatGPT ourselves?

You can, for simple tasks. Our clients come to us when the workflow touches money, customers or compliance, where "mostly works" is worse than manual, and when they want the saving measured, not guessed.

Do agents act without our approval?

No. Anything that spends money, contacts a customer or changes a system of record goes through approval gates you set. The consulting overview shows exactly where the gates sit.

What happens if we cancel?

Your workflows keep running; we hand over documentation and access. No data lock-in, same as our product.

Is our data safe?

Your data stays in your accounts wherever possible; everything else lives in a governed data layer with audit trails. We're approved API developers for Google, Meta and TikTok, which means we pass their security reviews.

Free AI opportunity assessment

Tell us where the hours go.

Three fields, no commitment. A Maaten engineer replies within one working day to arrange the assessment. In about a week you get the Automation Map: every workflow ranked by what it saves.

Name
Work email
Where does your team spend the most manual time?

Find out what the manual work is costing you.

The assessment is free and takes about a week. You get the Automation Map: every workflow ranked by what it saves. If we build, great. If we don't, you keep the numbers.

Get your free AI assessment

No long contract · Approval gates on everything · Measured from day one