Better decision workflows are the point. AI is an accelerant.
See the framework
How decisions improve
-
Work allocation
HumanToolAI
-
Decision pipeline
EvidenceRecommendationDecision
-
Information flow
NoiseClarityAction
The centre of gravity
Analytics-led decision workflows.
Start with the route from evidence to recommendation to decision. Then decide what stays human-owned, what belongs in deterministic tools, and where AI should support synthesis, framing, and translation.
Advisory
Collaborative work for analytics, reporting, forecasting, planning, and executive decision workflows.
Use the systemTools
Canvases, rubrics, workflow kits, prototypes, and small utilities that help teams operationalise analytics-led decision workflows.
Stay orientedSignal
Articles, briefs, field notes, and synthesis on AI, analytics, operating models, and decision quality.
How engagements work
Define proof before scale.
Every engagement starts from a recurring decision and names the evidence that would prove the workflow works — before anything is automated.
- Agentic weekly reporting
Separate evidence gathering from narrative generation, with a human owner verifying claims before stakeholders see them.
- Business context packs
Reusable, dated context with source hierarchy and known caveats, so analysts stop rebuilding the same background.
- AI readout quality control
A verification layer that tests claim-to-source links, confidence, and counter-evidence before publication.
- Workflow allocation
Map the workflow step by step with Human / Tool / AI, then name the owner of every consequential decision.
Latest signal · 1 October 2026
OpenAI’s DevDay shifts the emphasis from chat to persistent agents
Latest AI developments brief: current signal for leaders building practical AI systems and analytics-led decision workflows.