Frameworks · Research · Briefings · Field Notes
THINK DELUS
INTELLIGENCE
Here is how we think. Here is what we know. Here is the evidence. Decide for yourself.
How NGOs and Governments Prove Impact to a Skeptical Funder
A results report and a verified impact claim are not the same document. Funders and auditors are increasingly able to tell the difference, and treating them as interchangeable is how renewal conversations go badly.
The Retail Forecast Nobody Can Trace Back to a Source
A demand forecast that mixes traditional and modern trade data, promotions, and advertising spend is only useful if someone can explain why it says what it says. Most FMCG forecasting models can't.
What Evidence Governance Looks Like in a Smart City Rollout
A smart city roadmap is a governance document with sensors attached. Most rollouts get the sensors right and skip the governance - the roadmap looks finished on a slide and falls apart in year two.
When Unverified Data Is a Patient-Safety Risk
A fraud model with bad inputs produces a wrong number. A clinical decision-support tool with bad inputs produces a wrong treatment. Healthcare has the least tolerance of any sector for evidence that hasn't been verified.
Fraud Detection Needs an Audit Trail, Not Just a Model
A fraud model that flags a transaction without showing why is a liability in a regulated market. Banks, insurers, and fintechs need evidence a compliance officer can defend, not just a score.
What Regulators Actually Need to See: Evidence Standards for Utility Pricing
A pricing model that cannot show its metering data, loss calculations, and cost allocation logic will not survive regulatory review. Real-time telemetry only helps if it is governed as evidence, not just displayed as a dashboard.
What Institutional Evidence Looks Like in Agricultural Subsidy Programs
A subsidy program that cannot show which farmers actually received inputs, and what those inputs actually changed, is not a program an auditor can defend. Here is what evidence-grade subsidy data requires.
Institutional AI in East Africa
Accountability, data ownership, and measurable outcomes define successful public-sector AI. Learn the East Africa deployment reality.
Four Objections to Governance-First Digital Transformation — Answered
Africa needs speed, not process." "The pilot proved the technology works." "Governance is a Western import." "We can't afford it." Each objection sounds reasonable until it meets the evidence.
Why AI Makes Institutional Intelligence Urgent Now
Artificial intelligence does not fix weak institutional governance. It amplifies whatever governance already exists — good or bad. As AI adoption accelerates across African public sectors, governance has stopped being optional.
Dashboards Are Not Evidence
A dashboard is a presentation layer. Evidence is a property of what sits underneath it. Most institutions have invested heavily in the first and barely at all in the second.
How to Read an Evidence Table
Confidence levels, source types, and geography tags exist to stop a claim from being oversold. Here is how to read them, and why most published research skips this step entirely.
What "Evidence Lineage" Actually Means
Evidence lineage is a specific, checkable property, not a vague claim of data quality. Here is what it takes for a number on a dashboard to actually qualify as evidence.
Why 43% of Government AI Pilots Never Scale
OECD data shows 43% of government AI deployments remain stuck at pilot stage. The six institutional failure modes that keep African public-sector digital initiatives from reaching production.
What the Huduma Namba Case Teaches About Digital Identity Governance
A national identity initiative was halted by court order after roughly KSh 15 billion in spending. The lesson was never that digital identity failed — it was that institutional readiness determines whether digital infrastructure creates public value.
Where Does Your Institution Rank? The 5-Level Maturity Model
Most African public-sector institutions sit between Experimental and Repeatable on the TDII maturity model. Here is what each of the five levels actually looks like in practice, and what moves an institution up one.
The TDII Framework: How Institutions Turn Evidence Into Decisions
Institutional Intelligence is not a dashboard or a tool. It is a closed-loop framework for turning operational evidence into governed, repeatable decisions — this is how the TDII cycle works.
The grant classification error that kills African infrastructure bids
Project preparation and project implementation are different funding instruments. Confusing them costs institutions and their technical partners months of work, and funder credibility that doesn't come back.
Securing what you just digitized
Every meter, camera, or controller you connect to a network is also a new way in. Operational technology security is not an afterthought to instrumentation. It is part of the same design decision.
Before the dashboard, the sensor
Every real-time monitoring pitch starts with a beautiful interface. The actual work, and the actual cost, is the layer beneath it: sensors that survive dust and voltage spikes, and data that stays trustworthy after they do.
From match data to federation strategy
Proven evidence workflows from sports analytics apply to governance, procurement tracking, and impact measurement.
Why verified data beats more data
Governments and regulators need traceable evidence sets with audit trails. Learn why verified data beats more data.
Board-ready evidence standards
How Think Delus separates verified claims from opinion, and why that distinction determines whether a proposal advances.