Evidence-to-Decision Intelligence
Intelligence for institutions that must deliver.
Turn fragmented project, programme and portfolio evidence into earlier, traceable executive action — connecting institutional capability, strategic priorities, governance, delivery assurance, realised value and institutional learning.
One number governs this product: evidence-to-decision time. Design-partner pilots exist to measure it.
- Diagnose
- Prioritise
- Govern & Transform
- Monitor
- Assure
- Realise Value
- Learn
- Reassess
Prototype / Design-Partner Stage · Starting in Cameroon · Designed for Africa · English + Français
Evidence
Status report · risk register · steering minutes
Delivery signal
Evidence-linked contradiction / emerging issue
Governance implication
What leadership and governance must understand
Decision required
Traceable executive decision cue
Human review
ACCEPTAMENDREJECTESCALATE
No live model call. No real institutional evidence.
The evidence-to-decision gap
Projects do not fail in isolation. Institutions deliver them.
The problem is not a lack of reporting. It is the distance between evidence and decision.
- 01
Fragmented evidence
- 02
Late risk visibility
- 03
Slow executive decisions
- 04
Lost institutional learning
Evidence institutions already produce
- Status reports
- Risk registers
- Schedules
- Budgets
- Steering minutes
- Audits
- Evaluations
Institutions already produce large amounts of project, programme and portfolio evidence. The challenge is connecting that evidence early enough to capability, governance, intervention, realised value and institutional learning.
The institutional delivery loop
One institutional delivery loop — not eight disconnected tools.
Each stage produces evidence the next stage can use, and the loop closes back into institutional capability.
Select a stage to see the executive question it answers.
01DiagnoseExecutive question
Can this institution actually deliver what it has committed to?
Capability
CPLI™ · Institutional Capability Intelligence
Executive output
Capability baseline · improvement priorities · readiness context
02PrioritiseExecutive question
What should leadership fund, sequence or intervene in first?
Capability
MultiCriterIA™
Executive output
Transparent ranking · scenario comparison · decision trace
03Govern & TransformExecutive question
Who decides what, on what evidence, and through which structure?
Capability
Institutional Upgrade Roadmap™ · P3PMO / OPM
Executive output
Governance model · delivery-unit design · upgrade roadmap
04MonitorExecutive question
Where is delivery diverging from intent, and how early can we see it?
Capability
Delivery Intelligence
Executive output
Delivery signals · dependency indications · escalation candidates
05AssureExecutive question
Is this finding trustworthy enough to act on — and who validated it?
Capability
Delivery Assurance Intelligence
Executive output
Evidence-linked findings · assurance record · validated interventions
06Realise ValueExecutive question
Did the intervention create the promised value — and did it endure?
Capability
PPVSR™ · Value Realisation Ledger™
Executive output
Benefit-to-outcome trace · sustainability observations · value evidence
07LearnExecutive question
What can the next programme inherit instead of rediscovering?
Capability
Project Legacy · Africa Project Delivery Casebook™
Executive output
Institutional lessons · legacy records · comparable delivery patterns
08ReassessExecutive question
Has institutional capability actually improved since the last cycle?
Capability
Continuous institutional learning
Executive output
Updated capability view · revised priorities · compounding memory
Product proof
From evidence to a human-accountable decision.
One deterministic synthetic scenario across three institutional records. Connect the evidence and follow it through to the decision a named human must own.
Synthetic institutional records
R-1 · Monthly status report · Package 3
Civil works progress reported at 62% against a planned 74%.
R-2 · Risk register · Entry 118
Land-access risk raised from moderate to high; owner unchanged since last cycle.
R-3 · Steering minutes · Cycle 09
Mitigation action noted as ‘in progress’ for the second consecutive meeting.
Risk / dependency signal
—
Governance implication
—
Decision required
—
Why NPMTI compounds
The model is not the moat. The institutional learning system is.
- 01
Domain methods
An institutional-delivery doctrine that already exists as methodology and practice, not a generic AI wrapper.
CPLI™MultiCriterIA™Delivery AssurancePPVSR™ - 02
Governed institutional workflow
Evidence becomes human-reviewed decisions, decisions become interventions, and interventions are followed to outcomes.
- 03
Compounding institutional memory
Capability history, decision history, realised value and reusable project legacy accumulate inside the institution.
Value Realisation Ledger™Project Legacy
Defensibility is designed to compound as permissioned institutional evidence, decision history and post-project value records accumulate over time.
Where NPMTI fits
Built for a different level of the delivery problem.
Project / portfolio platforms
Primary object
Projects and portfolios
Typical role
Plan, coordinate and track delivery work.
Business intelligence
Primary object
Available operational data
Typical role
Visualise, report and analyse.
AI project copilots
Primary object
Project work, documents and user tasks
Typical role
Summarise, draft, search and assist decisions.
NPMTI Intelligence™
Primary object
The institution and its delivery system
Typical role
Connect institutional capability, prioritisation, governance, project and programme evidence, assurance, realised value and institutional learning.
Project and portfolio tools can remain part of the stack. NPMTI Intelligence is designed to add an institutional evidence-to-decision layer above and across them.
Project copilots can help professionals perform project work; NPMTI Intelligence is designed to help institutions improve the system through which that work is governed, assured and converted into enduring value.
This describes product categories, not named vendors.
Beyond project closure
Completion is not the same as value.
Did the intervention create the promised value — and did that value endure?
- Promise
- Delivery
- Outcome
- Sustainability
- PPVSR™
- Institutional lesson
- Reassessment
How it is held
Every project should leave the institution better able to deliver the next one.
Trust & responsible AI
Institutional evidence remains institutional evidence.
- Private by default
- Institution-controlled access
- Permission-aware disclosure
- Evidence traceability
- Human validation for consequential outputs
- No automatic public ranking
- De-identification or aggregation only where authorised
Private institutional evidence
Raw project evidence stays under institutional control and is not public by default.
De-identified benchmark intelligence
Comparative patterns only, and only where the institution has explicitly authorised de-identified use.
Institution-authorised public signals
Nothing appears publicly without explicit permission for that specific material.
NPMTI does not publish confidential institutional scores or rankings without explicit authorisation.
No security certification is claimed. Data-handling terms are agreed in writing for each engagement.
AI can support the decision. Humans own the accountability.
AI, integration and evaluation
AI is the reasoning layer — not the institutional judgment.
NPMTI Intelligence is being designed for material use of advanced AI, including Gemini for evidence-grounded reasoning and multilingual project-document understanding. Gemma is under evaluation for selected controlled deployment scenarios.
Institutional systems & documents
- PMIS
- ERP / Finance
- M&E
- Risk
- Schedules
- Reports
- Spreadsheets
- Collaboration records
NPMTI Intelligence™
- Evidence-grounded reasoning
- Cross-document synthesis
- English / French reasoning
- Risk and dependency signals
- Scenario explanation
- Executive briefing
Human validation
- Accept
- Amend
- Reject
- Escalate
Institutional action
- Executive decision
- Assurance
- Intervention
- Value evidence
- Institutional memory
These capabilities are in design and development. Connectors are planned, not released, and early work proceeds with exported or manually provided evidence under agreement. No model provider endorses, sponsors or is affiliated with NPMTI Intelligence.
Evaluation discipline
AI outputs must be useful, traceable and reviewable.
Evidence grounding
Can each material finding be traced to authorised evidence?
Decision usefulness
Does the output reduce ambiguity for an actual governance decision?
Human review quality
Can authorised users accept, amend, reject or escalate with clear rationale?
Evidence-to-decision time
Does the workflow materially reduce the time between usable evidence and an accountable decision?
These are the evaluation dimensions design-partner pilots exist to measure. No validated production performance is claimed, and no internal benchmarks, prompts or thresholds are published.
Use cases
Designed for institutions that carry delivery obligations.
Which investments are most ready to deliver — and where does leadership need to intervene?
- Compare candidate investments against agreed criteria with a reviewable decision trace.
- Surface delivery divergence across ministries and agencies before it reaches the political level.
- Follow financed commitments past completion into realised public value.
No client outcomes are claimed. These describe intended fit, not delivered engagements.
Product stage
Building with institutions — not pretending to be finished.
- NPMTI institutional-delivery architectureAvailable / defined
- CPLI™ methodologyAvailable / defined
- PPVSR™ / value methodologyAvailable / defined
- Project Legacy architectureAvailable / defined
- Trust and data principlesAvailable / defined
- Design-partner architectureAvailable / defined
- CPLI™ diagnostic prototypePrototype
- MultiCriterIA™ synthetic prototypePrototype
- Evidence-to-decision walkthroughPrototype
- Full four-view synthetic product LabPrototype
- Production AI reasoningPlanned
- Controlled deployment pathsPlanned
- Enterprise connectorsPlanned
- Longitudinal institutional intelligencePlanned
- Larger multi-institution deploymentsPlanned
Anything not listed as available or prototype is not offered today.
Prototype / Design-Partner Stage · Cameroon proof market · Designed for Africa
Commercial path
Start with one decision workflow. Expand into the institutional delivery system.
- Paid Pilot
- Institutional Subscription
- Premium Modules
- Enterprise / Multi-programme
- Network / National
Indicative early-commercial bands — not binding quotations
Design Partner / Pilot
Validate one high-value institutional decision workflow.
Scope
One scoped use case, agreed validation metrics and a documented evaluation.
USD 5k–20k platform + USD 5k–20k implementation
Institution
Establish recurring institutional delivery intelligence.
Scope
A single institution operating the delivery loop across its portfolio.
USD 25k–100k / year
Enterprise / Multi-programme
Connect multiple portfolios, governance workflows and premium intelligence capabilities.
Scope
Several programmes and delivery units under shared governance.
USD 100k–400k / year
Network / National
Scale governed intelligence across multiple institutions or a delivery network.
Scope
Cross-institution or national delivery systems.
USD 500k–2M+ / year
Funding routes
- Institution-funded
- Development-partner-funded
- Strategic sponsor-funded
- Network / group licence
Design-partner programme
Become a design partner
- 1Executive briefing
- 2Use-case definition
- 3Approved / synthetic / non-sensitive evidence
- 4Pilot configuration
- 5Validation metrics
- 6Pilot evaluation
- 7Scale decision
No file upload. No project evidence is transferred through this form at any point.
If a pilot progresses, an appropriate secure channel will be established for approved sample evidence.
FAQ
Questions institutions ask
What is NPMTI Intelligence?
An AI-native institutional delivery intelligence system. It connects institutional capability, strategic and investment priorities, project and programme evidence, governance, delivery assurance, realised value and institutional learning in one evidence-grounded intelligence system.
Is it a project-management system?
No. Project and portfolio management software plans, organises and tracks work. NPMTI Intelligence reasons across the evidence your delivery system already produces and connects it to capability, decisions, governance, value and learning.
Does it replace PMIS, ERP or BI?
No. It sits above them. Those systems remain the record of plans, transactions and reports; NPMTI Intelligence is the layer that connects their evidence into traceable institutional action.
How does MultiCriterIA™ fit?
MultiCriterIA™ is the decision and investment intelligence element. It makes prioritisation explicit — agreed criteria, applied weights, a visible ranking and a decision trace that can be reviewed later.
How does CPLI™ fit?
CPLI™ is the institutional capability intelligence element. It establishes the capability baseline that gives delivery evidence its context, and it is reassessed as real delivery evidence accumulates.
What does the AI do?
It is designed to reason over institutional evidence: long-context and cross-document reasoning, bilingual English/French understanding, risk and dependency extraction, evidence-linked synthesis, scenario explanation and executive briefing. These capabilities are in design and development.
What is human-reviewed?
Every consequential output. Findings and recommended interventions are routed for human validation — accept, amend, reject or escalate — before they become institutional record.
How is evidence protected?
Evidence is private by default, access is institution-controlled and permission-aware, outputs are traceable to source, and de-identified benchmark use or public disclosure requires explicit authorisation.
What is available today?
The institutional-delivery architecture, CPLI™ and PPVSR™ methodology, Project Legacy architecture, trust and data principles, the design-partner architecture, and interactive prototypes running on synthetic data.
What is still being built?
Production AI reasoning, controlled deployment paths, enterprise connectors, longitudinal institutional intelligence and larger multi-institution deployments.
What does a design-partner pilot involve?
An executive briefing, a defined use case, approved synthetic or non-sensitive evidence, pilot configuration, agreed validation metrics, evaluation, and a scale decision. Evidence sharing follows a separate written agreement.
Can it work in English and French?
Yes. Bilingual English/French operation is a design requirement, because the proof market is Cameroon and expansion runs through CEMAC and Francophone Africa.
Who is it designed for?
Government and public-investment bodies, enterprise and infrastructure owners, universities, and development-finance and programme institutions that must deliver across many projects at once.
Build the intelligence layer for better institutional delivery.
Design partners are being invited now. Prototype / design-partner stage — English and Français.