Method
The operating model AI transformation assumes you have.
AI does not fail on the technology. It fails on the model that decides what gets built, who signs, and what is kept. The failure is rarely a missing process — it is a missing record: the decision nobody can produce afterwards. Five parts make that model work — and all five are delivered here.
The five foundations
One model. Five controls.
Each foundation closes a different gap in the decision record. Together they make consequential change explainable and governable.
Gates something can actually fail.
Criteria are declared before commitment and every gate can produce a real no.
Demand & delivery →02 / AccountabilityA name against every signature.
Evidence, challenge and accepted risk stay attached to the accountable person.
Sign-off →03 / Process redesignOne front door. Nothing enters around it.
Every ask is compared against the same criteria, capacity and displacement.
Check gates →04 / DataDecisions held as data, not documents.
The proposal, challenge, change and signature form one durable trail.
Register & trail →05 / Workforce readinessAI drafts. A person decides.
Decision rights define where automation stops and human accountability begins.
Decision rights →Before and after
What actually changes.
The same six things happen either way. What changes is where each one is held — and whether it can be produced afterwards.
The portfolio before
The portfolio with MONO360
Demand
BeforeEmail, a spreadsheet, a business case someone already promised
With MONO360One governed front door
Approval
BeforeA meeting
With MONO360An explicit gate, with a name against it
Delivery
BeforeDisconnected systems
With MONO360One governed record
Reporting
BeforeManual consolidation
With MONO360Generated from the record
Risk
BeforeDiscovered in a review
With MONO360Surfaced against the record as it changes
Decision trail
BeforeDocuments
With MONO360Data that outlives the people who wrote it
01 · GOVERNANCE
Gates something can fail.
Control built into how work moves, rather than a review applied to it afterwards — the test is whether anything can actually be stopped. Governance is rebuilt as gates something can fail, set where the work is still cheap to change, with the refusals kept. A gate that has never stopped anything is renamed reporting.
02 · ACCOUNTABILITY
A name against every signature.
Responsibility that stays with a named person as capability scales, rather than dissolving into a process nobody owns. Every entry in the plan carries what it won its place against, who challenged it, what changed and who signed. Nothing is signed off by its author — including the designs behind the platform itself.
03 · PROCESS REDESIGN
One front door. Nothing enters around it.
The work reshaped around how decisions are actually made, rather than tools added to a process that already fails. One front door for demand, staged gates through delivery, and one record carrying the reasoning from strategy to build.
04 · DATA
Decisions held as data, not documents.
Decisions that can be compared, defended and reproduced — which requires the things they were made from to be held as data. The register, the capacity position and the decision trail are kept as governed data with access control enforced on every read, rather than as a folder of documents nothing joins. This is decision data, not an enterprise data estate: pipelines, quality programmes and the data platform itself stay with the organisation or its data partner.
05 · WORKFORCE READINESS
Who decides what, and where a person is required.
Whether people can actually operate the model — who decides what, where a person is required rather than merely expected, and how roles change once AI is working inside the loop. The operating model defines decision rights role by role, and the line between drafting and deciding is enforced rather than cultural: agents read, draft, score and challenge; a person signs, and that approval is enforced at the write seam in a single locked transaction, bound to exactly what was approved, rather than left to convention.
Fit
Who this is for—and who it is not.
The five foundations matter most where a decision left unmade at the front door is paid for later—in capital, delivery risk and consumed change capacity. Better to know before a call than during one.
MONO360 is for organisations where…
0120–100+ programmes compete for investment—with AI initiatives joining the queue unsequenced.
02Multiple business units own different parts of the portfolio, each with its own priorities and change demand.
03PMO data lives across multiple systems, and consequential investment cases must be proved from it.
04Excel, PowerPoint and source systems disagree, so nobody can say what was true when the decision was made.
05Governance happens after the decision has effectively been made—when intervention is already expensive.
06Executives do not trust the portfolio view they are asked to back new investment against.
07Nobody clearly owns the front door, so demand enters around the side.
Probably not for you if…
01The portfolio is small, or it is a single programme.
02What is needed is project tracking, not decision governance.
03An existing PMO already owns demand governance effectively.
04The requirement is another dashboard.
Illustration
How the loop runs.
Not a case study and not a client — the boundary described above, walked once end to end. The order is the point: the work is done before the decision, and the decision is still a person’s.
- Signal
- The register shows three programmes consuming capacity beyond their approved position while their reported status stays green.
- Agents
- Draft the exception — where the dependency exposure and the contention sit, and which gate it belongs at. They read, draft, score and challenge. They do not decide.
- Person
- The accountable role accepts it, rejects it, or sends it back for more. Nothing completes without that signature.
- Record
- The draft, the decision and the reason are kept together. The next gate reads them, and so does anyone who asks later why.
Where the gates sit
Gates fire before a line of code is written, not after.
A control that fires after the build is a report on money already spent.
In practice
How an engagement starts.
The five above are what a governed operating model needs. These are the two things you actually buy: a read of the portfolio you already have, and the software the model runs on.
ADVISORY
Reading a portfolio as it actually is.
The register, the capacity position, the last few quarters of decisions. What is in the plan that would not survive a gate, and what the gates that would have caught it look like. The work starts with what exists, not with what should have existed.
- Reads
- The register as maintained, not as presented
- Reads
- The capacity position against real availability
- Reads
- The last few quarters of decisions, and their basis
- Returns
- A written account of what would not survive a gate
- Returns
- A design for the gates that would have caught it
PLATFORM
The model, built as software.
Horizon PMO is a portfolio intelligence platform for enterprise change — one live view of every project, programme and investment. It exists because the method needed somewhere to live: a register that can be compared, gates that can be failed, and a decision trail that survives the people who wrote it.
- Views
- A suite of dashboards over one governed data model
- Tenancy
- Multi-tenant, with per-tenant isolation
- Identity
- Enterprise SSO and SCIM provisioning
- Access
- Row-level security policies in the database, with tenant scoping enforced on every query
- Scope
- Decision data — the register, the gates, the trail. Deliberately not your enterprise data estate
- Independence
- The advisory work does not depend on it
Start here
Start with the portfolio you already have.
Bring the plan as it stands. The first conversation is a read of it, not a pitch for any of them.
The first read ends in writing: which of the gaps are live in your portfolio, and the first gate that would expose them.