How a data and AI programme runs in Assettia
Assettia is the data and AI programme record: one place where an organisation keeps what its programme owns, what it has decided, and what it has proved.
Run data and AI as one programme, not a hundred initiatives. Evidence goes in with its source attached. Decisions are taken on the record, with the person and the reason. Value is tracked against baselines set when the money was committed. The eight workstreams below are the structure the programme is organised around, described in the same words the product uses.
First, the inputs
Before the workstreams, the evidence. Six modules gather what the programme needs to know about the organisation.
- Stakeholders: the directory of everyone consulted, interviewed or involved.
- Interviews: run executive interviews and draw the insight out of the transcripts.
- Surveys: send surveys and gather structured feedback from inside and outside the organisation.
- Templates: spreadsheet templates for gathering information in bulk.
- Business strategy: upload strategy documents; the platform draws out the objectives everything else links back to.
- Value benchmarks: evidence of what similar work returned elsewhere.
The eight workstreams
Each workstream is a set of modules that share a question. Nothing here is a separate tool: they read from and write to the same record.
Strategy
The direction: what the organisation wants from data and AI, and how it will know it is winning.
- Market signals: a curated feed of what is happening in the market and why it matters here.
- AI maturity: an honest assessment of where the organisation is starting from.
- AI strategy: one strategy record with three views: the vision, the themes drawn from the evidence, and the living board summary.
- Data strategy: the plan for the data the strategy depends on.
- Roadmap: the sequence: what happens now, next and later.
Value
The portfolio: finding, choosing, justifying and proving the work.
- Use cases: the library of every idea, from first thought to routed work.
- Prioritise: score and rank the use cases; agree the first wave.
- Evaluate: the case for each use case taken forward: a benefits case for internal opportunities, a business case for monetisation opportunities.
- Pilots: design the pilots that prove the value before scaling.
- Value tracking: committed value against what has actually been delivered.
Operating model
How the data and AI function is structured, governed and led.
- Operating model: how the data and AI function is structured, governed and led.
- Roles: the roles and job descriptions that make the model real.
People and adoption
Whether the organisation can actually use what gets built.
- Decision readiness: assess how ready people across the organisation are to use data in decisions.
- Customer focus: how customer-centred the organisation really is, measured.
Governance
The rules and the record: risk, policy, accountability. The risk register carries each AI risk with its controls and named attestations, and the regulatory obligations the organisation is subject to, tracked against its systems.
- Risk register: every AI risk, scored, owned and reviewed.
- Responsible AI: the organisation's responsible AI principles.
- Security policy: the AI security policy.
- Usage policy: the acceptable use policy for AI across the organisation.
Delivery
Turning decisions into working things.
- Delivery briefs: the brief that hands a use case to a delivery team with everything they need.
- Agentic studio: a safe place to design and trial agents.
Data
The foundations: what data exists, what state it is in, and who looks after it.
- Data requirements: what data the business questions actually need.
- Data assets: the register of the data the organisation holds, with rights and commercial position.
- Data quality: assess the quality of the data that matters most.
- Data readiness: whether the data is ready for AI to be built on it.
- MI blueprint: design what the organisation will report and measure.
- Data responsibilities: who does what for data quality across the organisation.
Monetisation
Treating data as a product with external customers.
- Propositions: shape data products the market might buy.
- Markets: the sectors and markets they could sell into.
- Business cases: the commercial case for each proposition.
- Client validation: test propositions with real prospects before building anything.
The executive front door
The Command Centre reads across every workstream in seven views: Overview, Programme, Strategy, Portfolio, People, Estate, Governance. Every figure names what it counts and opens to the records beneath it. Where a number cannot be evidenced, the platform does not show one. Board reports are produced from the same records.
The shape of the work
1
Frame the agenda. The programme opens on an eight-part agenda rather than a blank plan. Each workstream carries the activities, the decisions it must produce and the evidence it needs, so the first weeks are spent scoping the work rather than inventing the shape of it.
2
Gather the evidence. Interviews, surveys, assessments, strategy documents and data assets land in the record with their source attached. Nothing writes straight to a live record: everything inbound is proposed, reviewed and accepted.
3
Decide on the record. Use cases are scored, ranked and gated. Each decision names a person, a date and the evidence behind it, and revisits are recorded rather than overwriting what came before.
4
Commit and baseline. A funded case records its baseline at the go decision. Committed value and realised value are then two separate figures, and the gap between them is visible without anyone reconciling spreadsheets.
5
Run and review. Delivery briefs, pilots and workstream reviews sit against the same records. Off-track flags derive from the numbers, carry the owner's response, and cannot be quietly dismissed.
6
Report and hand over. Board reports are produced from the record. When the data leader changes, the successor reads the decisions, the rationale, the revisits and the evidence, and the programme keeps moving.
A realistic cadence
- The first weeks
- configuration, the first evidence pass, and the opening read on maturity and readiness.
- The first quarter
- the portfolio takes shape: use cases scored, the first wave agreed, benefits cases argued.
- Quarterly, then continuous
- strategy is revisited against the record instead of being rewritten from scratch, and the board report is a view rather than a project.
These describe how engagements are usually structured. They are not a commitment, and the platform never states a figure it cannot evidence.
The four rules
1
Deterministic first. Where a rule over records can answer, it does. Composed narrative is the exception, and it is labelled.
2
AI drafts, a person approves. Nothing generated writes to a live record. Output lands in a reviewable state and becomes a record when someone accepts it.
3
Every figure names its population. A count says what it counted. Click it and the platform shows the list.
4
Honest gaps. A register with nothing in it says so. An assessment not yet run is shown as not yet run, never as zero.
Questions people ask
Does an organisation have to use every module?
No. A programme uses the workstreams its questions need. Modules not in use show as empty, and the record says so rather than hiding them.
Where does a programme start?
With the inputs: stakeholders, interviews and the existing strategy documents. The first read on maturity and readiness follows from those, and the use case library starts filling from the first conversation.
What if the organisation already has a strategy?
It is uploaded as an input. The platform draws out the objectives, and everything that follows links back to them. The strategy is not rewritten; it is put on the record and revisited against evidence.
Can the organisation's own consultants run the programme in Assettia?
Yes. Consultants run engagements inside the client's record, and the client keeps the record when the engagement ends.
How soon does the board see a report?
As soon as there are records to report. An early report shows a programme that has captured much and proved little, which is an honest picture, and the point of having one.
See a programme run on the record.
A walkthrough takes forty minutes and uses a synthetic organisation, so nothing of yours is entered until you decide it should be.