Stop talking about AI. Ship it.
We embed with your team for 6–12 weeks: map the workflows, ship working agents, train your people, hand over the keys. Not a deck, not a "roadmap" — production-grade agentic systems deployed inside your operation.
From assessment to production agents.
We're not here to talk about AI strategy. We embed, ship production agents inside your workflows, train your team to operate them, and hand over the codebase. 6–12 weeks, not 6 quarters.
AI readiness assessment
Workflow mapping, task scoring, tech-stack audit. We tell you what to automate first — and what to leave alone.
Custom AI workflows
Agentic systems deployed inside your existing tools — Slack, HubSpot, Zendesk, internal apps. Your data, your rules, agents acting on top.
Chatbots & assistants
Internal assistants for your team, customer-facing chat for support, voice agents for ops. Built on your knowledge base, not a trained-elsewhere black box.
Data & analytics platform
If your data isn't ready for AI (and it usually isn't), we build the warehouse, pipelines, and retrieval systems that make agents actually work.
Team training
Your team learns to extend what we ship. Docs, playbooks, office hours. When we leave, you own the system — not us.
Governance & compliance
Audit logs, cost controls, PII handling, rate limits. Enterprise-grade guardrails from day one — not a bolt-on later.
How the ai transformation agent works.
A four-phase engagement designed to ship before the contract ends — not generate consulting deliverables.
Map + score
Weeks 1–2: interview your team, map current workflows, score every repetitive task by (automation ROI × risk × technical feasibility). Deliverable: a ranked automation roadmap.
Build working agents
Weeks 3–6: prototype the top 2–3 agents against real data, real tickets, real work. Quick cycles, honest about what works and what doesn't.
Ship to production
Weeks 7–10: deploy the winning agents inside your actual workflows. Monitoring, cost controls, fallback paths — all in place before traffic flows.
Train + document
Weeks 11–12: your team learns to operate, extend, and debug the system. Full code, full docs, full playbooks. We stay on call (fixed hours) for 60 days.
What you can order
from this engine.
Each deliverable has a scoped unit cost. Mix across services within your monthly allocation, or add 5-unit top-up blocks anytime.
| Deliverable | Type | Units |
|---|---|---|
| AI readiness auditWorkflow map + opportunity ranking + risk assessment | Research | 20= €2,000 |
| POC agent — single workflow3-week prototype against real data · go/no-go deliverable | Setup | 35= €3,500 |
| Production agent deploymentOne agent, prod-ready, with monitoring + guardrails | Setup | 60= €6,000 |
| RAG / knowledge pipelineEmbedding pipeline, retrieval, evaluation harness | Setup | 40= €4,000 |
| Internal AI assistantSlack / web chat · grounded on your docs | Setup | 50= €5,000 |
| Customer-facing chatbotSupport triage · routing to human fallback | Setup | 45= €4,500 |
| Team training programWorkshops + docs + office hours · 4 weeks | Consulting | 20= €2,000 |
| Strategy + governance retainerMonthly review · cost, safety, effectiveness | Consulting | 10= €1,000 |
Unit estimates are tentative — final scope is set during your first sync. Because agents compress the work, you often get more output per unit than a traditional hourly retainer would predict.
From kickoff to first production-deployed agent handling real work. Typical engagement lands in 8–10 weeks.
Across workflows we've shipped agents into. Measured pre/post with matched task samples.
Across our engagements. Support triage, proposal drafting, lead research, content generation, internal Q&A.
Fixed on-call window after handover. Bug fixes, coaching, minor extensions included.
Three shapes where AI transformation returns hardest.
AI isn't magic — it's a leverage tool. Certain company shapes get outsized returns. Others are better served by the other 8 services.
Services businesses with repeatable work
Agencies, consultancies, professional services. Proposal drafting, research, onboarding, reporting — the 40% of time spent on boilerplate.
- Billable-hour model
- Repeatable deliverables
- Scaling bottleneck
Teams drowning in support tickets
SaaS / ecom / services with growing support volume. Triage agents, response drafting, knowledge-base grounding — 30–50% of ticket load handled.
- Growing ticket volume
- KB exists
- Hiring isn't the answer
Companies with deep internal knowledge
Law firms, healthcare, complex B2B. Internal assistants that know your docs, your process, your history — and don't hallucinate outside them.
- Tribal knowledge
- New-hire ramp pain
- Compliance matters
AI Transformation lives inside your monthly unit allocation.
Typical ai transformation engagement: Full engagement: 20 units audit + 60 units production + 20 units training = 100 units over 10 weeks. That fits comfortably in Liftoff — the tier most clients pick.
AI Transformation, specifically.
The questions we hear most often about this engine. Don't see yours? Ask us.
Is this just GPT wrappers with fancy branding?
No. We use foundation models where they fit and build domain-specific layers on top: retrieval, evaluation harnesses, guardrails, fallbacks, cost controls. A production agent is 10% model call, 90% surrounding system. The surrounding system is where we add value.
What about data privacy — we can't send stuff to OpenAI?
We support on-prem and private-cloud deployments (Anthropic Bedrock, Azure OpenAI, open-weights models on your own infra). Zero-retention agreements are standard. For regulated industries we've shipped fully-air-gapped setups — it's slower, but possible.
How do you handle AI hallucinations / wrong answers?
Every agent has an evaluation harness — a test suite of real inputs with known-good outputs — that runs on every model change. We ship confidence thresholds; low-confidence outputs route to humans. No silent failures.
What if we want to keep extending after you leave?
That's the goal. We train your team, document the system top-to-bottom, and leave a playbook for extending it. The codebase is yours. Post-handover we offer a 60-day fixed on-call window, then optional monthly retainer for ongoing work.
How do you price something this bespoke?
Unit-based, same as the rest. A typical 10-week transformation lands at 100–150 units (€10k–€15k). Pure technical cost, no consultancy-ware markup. We'll scope honestly in the audit — if your highest-value task can't be agent-automated safely, we'll tell you.
Other engines that fire well with this one.
These services share data and playbooks with AI Transformation — stacking them inside the same package compounds results faster than running them in silos.
Ready to fire the ai transformation
engine?
Book a 30-minute conversation. We'll be honest about whether agentic transformation fits where you are today.