AI IMPLEMENTATION · SME ROADMAP
AI Implementation for SMEs: A Roadmap That Ends With Working Systems
Not a strategy document. Not a pilot with no follow-through. A 6-8 week engagement that ends with AI running in your core processes and your team capable of maintaining it.
The failure modes I see constantly
-
✕
AI pilot ran well in the sandbox. Died at the point of integration with the real system. Nobody owned the handoff.
-
✕
Vendor sold a solution before understanding the problem. You're six months in and the use case doesn't fit your actual workflow.
-
✕
Implementation finished. Vendor left. Your team doesn't know how to use it, maintain it, or extend it.
-
✕
AI budget consumed. Board wants to see results. The slide says productivity up 12% — nobody can explain how that was measured.
The 6-8 week implementation structure
- 01
Discovery and scoping (week 1)
Map the 3 highest-ROI processes for AI intervention. Define success metrics before writing a line of code or signing a vendor contract.
- 02
Stack selection (week 2)
Evaluate models, APIs, and infrastructure against your actual constraints: budget, data residency, security, and team skill level.
- 03
Build (weeks 3-6)
Implement the AI layer in production — not a demo environment. Your team is part of the build, not watching from the side.
- 04
Validation and measurement (week 7)
Run against the success metrics from week 1. If the number isn't there, we extend. No vanity metrics.
- 05
Handoff and documentation (week 8)
Full documentation, runbooks, and a 30-day hypercare period. You should be self-sufficient on day 60.
Proof-of-work
-
Frihet ERP — AI-native ERP implemented from zero to production in 8 months. 40+ AI tools embedded, used daily by paying customers.
-
Firma — AI implementation for a 30-lawyer legaltech platform. Booking agent, document automation, and client portal in production.
-
ICEN Academy — 13 master's programmes launched with AI-assisted content, payment automation, and student management.
Pricing
AI Implementation Roadmap (6-8 weeks): from 15,000 EUR. Scope depends on complexity and number of processes. Fixed price, not T&M.
FAQ
- What sectors do you work with?
- SaaS B2B, legal and consulting firms, fintech, manufacturing, and e-commerce. The common thread is mid-market companies with complex internal processes.
- Do we need an internal tech team?
- Helpful but not required. I've implemented AI for companies with zero technical staff. The handoff approach adjusts to your team's capability.
- Is the price fixed or variable?
- Fixed price based on agreed scope. No surprise invoices. If the scope changes, we agree a new price before proceeding.
- What if the implementation doesn't work?
- Defined success metrics upfront mean there's no ambiguity. If metrics aren't met at week 7, we extend without additional charge.
- Can you implement AI without replacing existing systems?
- Almost always. AI layers sit on top of existing infrastructure. Rip-and-replace is rarely the right answer — and never the first one.
Go deeper
-
Magazine
AI Operating Models 2026 — the 5 patterns that scale →
-
Magazine
Agent orchestration 2026 — the false sense of scale →
-
Magazine
Zero-click operations — design for teams that scale →
-
Services
Systems Sprints — 2-4 week implementation →
-
Training
In-house programs for executive teams →
-
Diagnostic
Free 15-min AI diagnostic →
Six weeks from now, AI is in production.
Or it isn't, and we haven't started. Let's find out in 30 minutes.
Book a scoping call