Plain writing about what's actually happening with AI inside Pakistani institutions. Mostly observations. Occasional opinions. No whitepapers in lieu of working systems.
Awareness is at 76 per cent. Adoption is under 10. The gap between subscription and capability is where most institutional AI budgets are dying right now. A breakdown of where the money is going, and why nine in ten of those deployments will be quietly shelved by Q4.
The ministry scored 22. They expected 60. The five-dimension breakdown explains why — and what changes when you stop scoring AI investment by spend and start scoring it by capability lift.
A pilot built right. Champions trained right. Adoption collapsed inside six weeks of go-live. The reasons are boring, structural, and almost entirely solvable. Why no Pakistani AI firm publishes the Embed pillar as a packaged SKU — and what that means for buyers.
Fractal IPO'd. Mu Sigma capped at $1B. Wadhwani embedded in ministries. Tiger Analytics married Microsoft. Four founder-defining decisions, four founder-defining mistakes — and the four that map cleanly onto Pakistan in 2026.
International quality at Pakistani cost structure is a positioning, not a discount. The pricing bands, the margin logic, and the engagements we refuse because the unit economics never work.
PC-1 fluency. PPRA compliance. Vendor empanelment. SIFC pathways. The procedural ground game that decides who wins the USD 1B in committed AI spend.