The Intelligence Delivery Life Cycle
How we design and build integrated intelligence: in phases, each one ending with your decision to go on, change course or stop.
The software is how the intelligence reaches the work.
What you are buying is a decision or a task done better, by a person or an agent, inside an operation that has changed to make room for it.
A software life cycle ends when the software is deployed. This one has more to get right: the decision, the data behind it, the people who use it and the evidence that it works. So the measure and where it stands today are agreed before anything is shown, and every release shows the value against it.
Commit in steps, each funded on evidence.
Focus & Align settles where to start. Then each of four phases answers one question, and ends with your decision to go on, change course or stop. Either side can stop at any gate, and stopping early is a good result: it saves the cost of building the wrong thing.
-
Scoping and Specification
Is this worth pursuing, and what should we build?
- What you see
- Demonstrators, then prototypes on extracts of your data, in our environment.
- How it’s judged
- Agreeing with your experts what good looks like.
It ends with A quantified business case and a specification.
-
Controlled Trial
Does it work in a limited version of your world?
- What you see
- Alphas in your environment, on real data, with a small group of experts.
- How it’s judged
- A set of real cases, judged by your experts.
It ends with A decision to fund the build, at a firm price.
-
Operational Pilot
Does it hold up with real people doing real work?
- What you see
- Betas in production, with a bounded group of users.
- How it’s judged
- A benchmark in production.
It ends with Security and user acceptance signed off, and performance against an agreed floor.
-
Rollout
Does it run as part of how you work?
- What you see
- General availability, for everyone it was built for.
- How it’s judged
- Continuously, with drift watched for.
It ends with Handover into Continuous Intelligence Operations, with runbooks and drift monitoring.
Reach widens as confidence is earned.
Something runs at every release, so there is never a big reveal. Each one shows working software, data analysis, decision analysis and the value against the baseline, and more people and more real data come in only as each release earns it.
| Release | Where it runs | On what data | What we stand behind |
|---|---|---|---|
| Demonstrator | Our environment | Your own example | The idea, made visible. Always thrown away. |
| Prototype | Our environment | Extracts of your data, once a data-protection decision is recorded | The findings and the specification. Thrown away too. |
| Alpha | Your environment, under control | Your real systems and data, as far as a fair trial needs | A fair test, with the options compared |
| Beta | Your production environment | Live production data | The software, against a floor your experts agreed |
| General availability | Your production environment, fully instrumented | Live production data, watched for drift | The service levels in the contract, measured and reported |
The price firms up as the unknowns close.
Early on, the least is known about your data, your systems and how decisions really get made, so that work is time-boxed. Once a trial in your own environment has settled those questions, the build has a firm price.
Rollout hands over a loop that is already turning.
From the first Alpha, two kinds of work run alongside the build.
- Embed
- Your people are trained on the system as it grows, with the decision governance and new ways of working it needs.
- Continuous Intelligence Operations
- Where people and the system disagree is recorded from the first Alpha, and used to improve it. By rollout, it is how the system runs.