Hakkoda, an IBM Company / Leadership briefing
One source of truth. Every build we can point at.
157 client engagements in one structured library. Fifteen assets built on it. Every claim traces back to a record, and the records carry their own evidence grade, so nothing reaches a buyer on a number we cannot stand behind.
Internal use only
Figures on this page are counted from source_of_truth_compiled.json as of 15 September 2026. No client name here is approved for external use without confirmation.
01
What this is.
An alternative to one-off, deadline-driven collateral. A content engine sitting above a structured library of client proof, mapped to capabilities and verticals.
The output is sales and marketing material that is faster to produce, consistent across teams, and defensible in terms of data accuracy.
Every record is extracted from raw source material. Win-story decks, the master case-study deck, engineer write-ups, and the public site, then normalized into one schema: capability, vertical, delivering team, architecture, metric, and evidence grade. Unstructured use cases become queryable, comparable data.
A living source of truth that compounds in quality and reuse with every engagement. It feeds off-the-shelf assets and custom collateral built on the fly for a specific pursuit.
Every figure has to trace to a validated record, and the asset has to hold to IBM brand standards. The templates enforce the brand side: a shell that fails the lock check, or a client not in the library, will not render. Data accuracy is a human review gate, applied as new records enter.
02
The library.
Each record carries an evidence grade taken from the source, not assigned by us.
Hard metric
90
A confirmed, quantified outcome verified in a source. These lead an asset and carry the headline number.
Structural
46
Named client, real architecture, confirmed build, no percentage on file. Credible without a number, and never given one.
Directional
16
Projected or estimated. Labeled as such wherever it appears, never presented as a realized result.
Unconfirmed
5
Withheld. Nothing client-facing is built on it.
69 records are Hakkoda-delivered, 62 IBM Consulting, 16 other IBM teams, 9 blended or partner-delivered, and 1 unresolved.
Seven capabilities and fourteen verticals. The deepest are Healthcare at 33, Retail and CPG at 25, Financial Services at 22, and Life Sciences at 10. Life Sciences is kept separate from Healthcare on purpose.
03
The templates.
Two shapes. Both Carbon-compliant, both co-branded, both locked against off-brand drift.
Three-pager
Seam
Linear brief
The job
Present proof
A fixed Triptych. Outcomes, systems, success. The proof table is the point.
The job
Make an argument
A single scroll with a contents rail. Six to eleven sections. The argument is the point.
Enforcement
Hash-locked shell
A lock check and a linter run before deploy. A fabricated client or a mislabeled delivery does not render.
Enforcement
Ten components
Seam, map, flow, proof index, status ledger, split, architecture, close.
04
Fifteen assets.
Newest first. Placeholder means the shell exists and the content does not.
Proof libraryAll 157 records. Five filter axes, and it generates a standalone sub-asset from any filter state.
Interactive
Tool
FleetBuilt for ourselves, packaged to sell.
Linear brief
Internal build
Project management agent (Medtronic)Cross-region agent execution. Internal only, attribution unresolved.
Linear brief
Client build
Built on your stackIntegrated with the technology a client already runs. 17 Hakkoda rows, 8 IBM Consulting.
Linear brief
Capability
What we have built (Pfizer)Nine clients across semantic layer, governance, and AI readiness. All blinded except UFC.
Slide
Client deliverable
GTM leadership brief (v1)This page. Replaces the August version.
Linear brief
Leadership
Healthcare
Three-pager
Vertical
Retail & CPG
Three-pager
Vertical
Snowflake
Three-pager
Partner
Agentic Orchestration
Three-pager
Capability
Trusted Data and Governance
Three-pager
Capability
Intelligent Automation
Three-pager
Capability
Decision Intelligence
Three-pager
Capability
Data Modernization
Three-pager
Capability
Customer & Product Intelligence
Three-pager
Capability
Strategic advisoryShell only. Every field is still a placeholder. Not built.
POV
Placeholder
—
05
Assets on demand.
Stack filters across capability, theme, vertical, delivering team and evidence grade, then generate. A standalone page opens with the matching records, sorted strongest evidence first and titled from the filter state.
Integration plus Healthcare returns ten records in three clicks, five of them with a hard metric. The page is ready to save, send or deploy.
It produces the evidence half of an asset. The headline, the argument and the architecture diagram still come from a person.
06
Where it is thin.
The same pass that finds proof also finds its absence.
GapNo disclosure field.Nothing in the library records which of the 157 clients can be named externally. Until that exists, every asset carries a blanket warning.
GapTen hybrid records.Graded Hard metric because the source leads with Confirmed, but each carries a directional caveat inside. Check before any goes into a headline.
GapFleet and the agent builds have no client record.Neither has a confirmed outcome metric. Both carry status ledgers instead of proof tables.
ThinGovernance in healthcare.Ten records, two with a hard metric. Enough to talk about, not enough to lead with.
ThinAgentic Orchestration has no Hakkoda spine.Seventeen records, strong on the IBM Consulting side, nothing Hakkoda-delivered to anchor it.
07
What is next.
Add a disclosure field.
157 decisions on which clients can be named. Nothing else stands between the tool and a client deck.
Close the thin spots.
The library now tells us what to collect next. Agentic Orchestration needs a Hakkoda build. Governance in healthcare needs a number.
Open it to the team.
An access-gated data room where anyone can filter and generate their own cut, on the same locked templates, from the same validated records.
Counted from source_of_truth_compiled.json, 157 records, 15 September 2026. Evidence grades follow the leading term of each record's own metric status, so the four buckets sum to 157 exactly.
Verticals normalized from 75 raw values to 14, delivery from 34 to 5. Normalization runs in the build script. The figures regenerate on every rebuild.