An audit that tells you what to automate first — and what to leave alone
We compared 227 documented types of editorial AI use against a real production platform. You get the same comparison for your newsroom — with what you can switch on tomorrow separated from what needs building.
227 use case types, drawn from documented practice
The taxonomy is built on the J·Index catalogue — approximately 2,989 documented implementations across 1,807 newsrooms in 140 countries. It is not a list of what vendors advertise. It is a list of what newsrooms actually did.
Snapshot taken 26 August 2026. A one-off snapshot, not a live index.
Three axes, no opinions
Each of the 227 types is classified on three axes:
Architecture
present / partial / absent. Does an existing step or route configuration close it?
Route to it
a step or a setting. Does it need code, or is a prompt and a configuration enough?
Actual use
running / occasional / no. Real run volume in production, not a hypothetical capability.
Every mark is backed by a concrete step, a concrete route and a run count. Not by "we think we could do that".
The result, including the uncomfortable part
- Running in production every day
- 26% · 59 of 227
- Technically available, but idle
- 35% · 80 of 227
- Not in the architecture
- 39% · 88 of 227
- Technically available in total
- 61% of documented world practice
- 26%Running in production every day
- 35%Technically available, but idle
- 39%Not in the architecture
Snapshot taken 26 August 2026. A one-off snapshot, not a live index.
The gap between "can" and "do" is the point. In our own case 61% was technically available and 26% was actually running. That gap is what the audit looks for in your newsroom. We have measured one newsroom so far — our own — and we do not pretend otherwise. Most of the missing cases sit in areas we deliberately do not build: reader personalisation, community and moderation, audience analytics, reader-facing search and chatbots, monetisation, litigation. Genix ends at publication. A content production tool should not also try to be a CMS, an ad stack and a comment moderation system. This is an internal audit by the Genix team, using a public taxonomy. It is not third-party research and we do not present it as one.
The deliverable
A coverage map
of your editorial process across the 227 types, on the same three axes.
The "switch on" list
what your team can run with a prompt and a configuration, no development.
The "needs integration" list
with the specific dependency named, not "requires integration work".
The "not worth it" list
including the things we deliberately do not build, so you can stop looking for them here.
A recommended pilot scenario
with the reason it was chosen and the measurable outcome it should produce.
Twenty-eight cases that need a prompt, not a release
In our own audit, 28 cases were already supported by the architecture but not used. Fifteen of those did not depend on any external data source at all.
Interview questions from a speaker dossier · quizzes and polls attached to a piece · a policy-compliance gate before publication · OCR of scanned documents · object recognition in photos · classification of manipulative content · prioritising claims for fact-checking · storyboards · subtitling in other languages · transcription in low-resource languages · SEO keyword selection.
This is the difference between a capability and a habit. The audit finds the capability you have already paid for and are not using.
How the audit runs
A paid, standalone first engagement.
It is useful whether or not you go further.
Price.
Agreed individually and invoiced by bank transfer.
Before it:
a free 30-minute qualifying call, so neither side buys an audit that should not be bought.
Setting expectations
It is not a free consultation. It has a deliverable and a price.
It is not independent third-party research. It is our method, applied by us.
It is not a product demo. If you want to see the platform, that is a different meeting.
It does not conclude with "buy the platform". A number of findings in any audit are things you should fix without us.
Find out what you already have
Book the qualifying call. Thirty minutes, free, and we will tell you whether the audit is worth buying for a newsroom your size.
Book a call