“How long will this take?” is a fair question — and the honest answer is “it depends, but less than you fear for a focused use case.” Here’s a realistic view for 2026. (dgm implements osFoundry as an independent partner.)
The realistic shape
- A focused pilot: often weeks, not months.
- Full integration: depends on data and system complexity.
- Scaling across the business: iterative, over months as you add use cases.
The key insight: the AI model is rarely the bottleneck. Data preparation and integration are.
What drives the timeline
| Driver | Effect |
|---|---|
| Data readiness | Messy/scattered data adds time |
| Number of systems | Each integration adds work |
| Governance needs | Compliance-heavy contexts take longer |
| Decision speed | Slow internal sign-off slows everything |
Notice that two of these (decision speed, governance readiness) are about your organisation, not the technology.
How to go faster
- Start with one use case, not a transformation.
- Use data you already have, in good-enough shape.
- Avoid scope creep — the classic timeline-killer.
- Use a pre-built, usage-priced platform rather than building from scratch.
- Use retrieval over existing documents instead of long data-engineering projects.
A simple internal use case (document handling, knowledge retrieval, drafting) can often show value in weeks.
Why phase it
Phasing beats a big-bang launch: deliver one use case, prove it, learn, then expand. You see value early, de-risk the project, and build momentum and trust — rather than waiting months for a launch that might miss.
Where osFoundry and dgm fit
dgm runs phased delivery on osFoundry: a quick pilot first, then integration and scaling. Because osFoundry is pre-built and usage-priced (not a from-scratch build) and supports retrieval over your existing documents, deployment of a focused use case is fast. Self-hosting or an EU region handles UK data control (it publishes US/EU/JP regions, not a UK one). We assess your data and systems up front so the timeline is realistic, not optimistic.
dgm is an independent integration partner with zero integrations so far. To get a realistic timeline for your use case, book a consultation with dgm. General information, not specific advice.