Case Studies

Six engagements where we deployed agents —
and then made ourselves unnecessary.

Every case in this portfolio follows the same pattern. We deploy the first wave of production agents ourselves. We train your staff to use Claude Code or OpenClaw at the same level our consultants do. They build the next wave. We exit on schedule.

6
Engagements
53
Agents shipped
217
Staff trained
€36M+
Annual value created
01 · Not a product
No SaaS subscription, no software license, no recurring per-seat fee.
Every agent runs on the client's own infrastructure (or their own cloud tenant), against models the client controls. Zero vendor lock-in.
02 · Not a project
No black box, no permanent maintenance retainer, no "only we can change anything."
We deploy the first wave fast. Your staff own everything we built and can extend it without us. We are not the bottleneck on your roadmap.
03 · A capability
A 12–14 week Academy that turns staff into agent operators and engineers.
By the end of every engagement the client has 8–25 internal operators, 3–6 internal engineers, and a leadership team that no longer needs to translate AI through us.
the portfolio

Six clients. Six departments. One method.

From the World Health Organization in crisis to a 14-person research lab in Prague. The shape of the engagement is the same. The outcomes are durable because the capability stays with the client.

Public sector International org Claude Code
$1.2M – $2.0M/ year saved
15 production agents · 87 staff trained · 50% workforce reduction absorbed

Teaching the WHO to build: from crisis to self-sustaining AI engineering

When the US withdrew funding and WHO had to cut nearly half its 7,400-person workforce, we deployed eight production agents in 16 weeks — and trained 87 WHO staff to build the next seven themselves. The Head of HR personally recommended the program to the Director-General.

Workforce held at80–90% capacity
Staff trained87 graduates
Engagement18 months
Read the case
Public-sector consulting EU procurement OpenClaw · sovereign
+€3.6Min won contracts
Win rate 22% → 38% · Procedural disqualifications cut 95% · 9 agents deployed

Winning the tender: how a 340-person government advisory firm out-bid firms 10× its size

Tender response is a volume game and a depth game at the same time. We deployed five agents and trained 38 consultants to build the next four — all running on the firm's own infrastructure to satisfy the strictest confidentiality clauses.

Throughput per consultant+383%
Bid prod cost−71%
Engagement9 months
Read the case
Marketing Performance agency Claude Code
34% → 61%gross margin
6.1× campaign output · 23 → 41 clients · zero senior hires

From agency to operator: 6× output, almost 2× margin, no new senior hires

A 47-person performance marketing agency deployed seven agents on Claude Code and trained 22 strategists and creatives to build the next four. Campaigns ship in 3 days instead of 9, account managers carry 12 clients instead of 5, and the partners run a business now instead of operating one.

Revenue+71%
Time-to-launch9d → 3d
Engagement9 months
Read the case
HR & talent Executive search OpenClaw · sovereign
18 days → 4to shortlist
Placements per consultant 2.2× · PE win rate 27% → 41% · +€1.2M placement fees

From 18 days to 3: better shortlists in a fraction of the time

A 62-person executive search firm placing CFOs, COOs, and CHROs deployed six agents on OpenClaw — running entirely on the firm's own infrastructure to honour candidate confidentiality — and trained 19 consultants and researchers to extend them.

Universe coverage200 → 600 candidates
Mandate-to-placement71% → 84%
Engagement8 months
Read the case
Sales Industrial SaaS Claude Code
+€2.4Mnet new ARR
Meetings booked 2.6× · SQL → close 11% → 16% · 52% fewer outbound emails

From outbound to outcomes: 2.6× meetings, 73% lift on close rates, no new sales hires

A 78-person commercial team at a Central European industrial-SaaS company deployed eight agents on Claude Code and trained 31 reps and RevOps engineers to build four more. They send fewer emails and book more meetings. The CRO's three-year ARR plan is ahead of schedule.

Forecast accuracy±18% → ±6%
Sales cycle142d → 98d
Engagement9 months
Read the case
the pattern

The same three phases work in every
department, sector, and language.

What changes between engagements is the shape of the agents. What does not change is the shape of the engagement: deploy fast, transfer the discipline, exit when the client can keep going alone.

01 · Ignite
First production agent in your environment within days
Pick the most painful workflow in the most painful department. Ship a real agent into production. Prove the model on real work, real data, and real users. No demoware.
02 · Pilot & Academy
Department-scale rollout while we train the first cohort
3–5 agents shipped by us in parallel with a 12–14 week Academy that takes your staff from operator basics through workflow design, tool-building, evaluation, and governance. Cohort 1 graduates ship their own first agent before we leave the room.
03 · Scale & Transfer
Your team builds the next wave. We move toward zero footprint.
Internal builds outpace external builds. The Academy starts running itself, taught by your own graduates. By exit, you have an in-house "agent engineering" function and we are no longer required to keep the program running.

Want to be the seventh case study?

A 60-minute discovery call. No software pitch. We map your most painful workflow, scope a first agent, and tell you honestly whether this engagement model fits your organization. If we cannot find value in 7 days, you pay nothing.