Editorial Intelligence OS
The libraries, workflows and AI-assisted processes that operationalise Editorial Intelligence
Documents the practical infrastructure of Editorial Intelligence — the prompt libraries, signal capture workflows, narrative management processes and AI-assisted tools that turn methodology into operational reality inside an organisation.
Version 0.1 — in development. This experiment documents a system being built in parallel with its own documentation. The structure is stable; the individual components are being added as they are tested.
Purpose
The Editorial Intelligence frameworks describe how to think. The narratives demonstrate what the thinking produces.
This experiment documents the layer in between: how the methodology actually runs inside an organisation — the specific tools, workflows, prompts, templates and processes that make Editorial Intelligence operational rather than theoretical.
The Editorial Intelligence OS is not a product. It is a working system — built for a specific context, documented honestly, and made available as a model for organisations building their own version.
The problem
Methodologies fail at implementation.
A team reads a framework and understands it intellectually. They want to apply it. But the gap between understanding a methodology and having the operational infrastructure to run it — the templates, the prompts, the workflows, the meeting rhythms, the decision points — is where adoption breaks down.
Most organisations that try to implement an editorial intelligence approach build the same infrastructure independently, from scratch, making the same mistakes. The knowledge of what works and what doesn’t exists inside individual organisations but is rarely shared.
This experiment is an attempt to change that: to document the operational layer of Editorial Intelligence in enough detail that another organisation could adapt and implement it.
What the OS contains
The Editorial Intelligence OS is organised into four operational layers:
Signal capture The tools and practices for collecting signals from multiple sources — customer conversations, research outputs, event intelligence, regulatory changes, internal data — and routing them into a shared system where they can be connected.
This includes: signal capture templates, tagging conventions, source prioritisation, the weekly signal review process, and the criteria for escalating a signal to a narrative team.
Narrative management The workflow for developing, maintaining and evolving narratives over time — from the initial brief to the first version, through the evidence-gathering cycles that strengthen the position, to the decision about when a narrative has run its course.
This includes: narrative brief templates, version control conventions, the evidence review process, cross-narrative connection mapping, and the criteria for retiring or merging narratives.
AI-assisted workflows The specific prompt libraries and AI workflows that accelerate signal processing, narrative drafting, evidence packaging and connection identification.
This includes: the Case Study Agent prompt library, the signal synthesis workflow, the narrative brief generator, the cross-connection identifier, and the editorial review checklist for AI-assisted outputs.
Editorial governance The decision-making processes that maintain quality and strategic coherence as the system scales — who has authority over narrative direction, how conflicting signals are resolved, how AI outputs are reviewed, and how the system learns from what doesn’t work.
Approach
The OS is being built alongside the methodology documentation rather than after it.
This is deliberate. A methodology documented in isolation from its implementation tends to become idealised — it describes the system as it should work rather than as it does work. Building and documenting simultaneously means the OS reflects operational reality: the workarounds, the edge cases, the processes that had to be simplified because the elegant version created too much friction.
Each component of the OS is documented in three parts:
- What it is — the tool, template or workflow
- Why it exists — the specific problem it solves or the gap it closes
- What we learned — what changed between the first version and the current one, and why
Current status
The signal capture layer is the most developed. The templates and tagging conventions are in use and have been iterated based on three months of operation.
The narrative management workflow is partially documented. The brief templates and version control conventions are stable. The evidence review process is being refined.
The AI-assisted workflow library is growing in parallel with the Case Study Agent and Hidden Hours Diagnostic experiments. Each experiment produces prompt templates that are incorporated into the OS library.
The editorial governance layer is the least developed — partly because it requires the most organisational context to implement, and partly because the right governance model depends on the scale and structure of the team using the system.
Next iteration
- Complete the signal capture documentation with worked examples from the Hidden Hours and Winning in Small signal gathering process
- Publish the first version of the prompt library as a downloadable resource
- Document the narrative brief template in full, with a worked example from the Trusted AI narrative development process
- Begin the editorial governance section with a focus on AI review processes — given the volume of AI-assisted work now in the system, this is the most urgent governance question