Announcing collaboration with Oxford AI, to build the world's most accurate industry forecasting platform.

Early results from an ongoing study

Adding forecasts produced 98% more AI citations.

We compared 50 plain articles with 50 similar articles that included an Authority forecast, fresh data and expert commentary. After five days, the articles with Authority were cited almost twice as often by AI search.

98%

more AI citations

70%

more individual articles cited

5 days

to produce the early result

What we did

We compared the same kind of content, with one clear difference.

Both groups covered valuable, competitive business topics. One group stayed as normal blog content. The other group received a live Authority forecast, original data and commentary from named experts. We then tracked which articles AI search engines cited when answering relevant questions.

Plain article50 articles

Industry guide

How new AI rules could affect employers

Useful, but similar to many other search results.
Article with Authority50 articles

Industry guide

How new AI rules could affect employers

Live Authority forecast

Updated daily

Will new employer AI rules be introduced by September 2027?

ProbabilityNew evidenceExpert view
Same useful article, plus information no other page has.

100 similar articles

50 plain articles and 50 articles with Authority forecasts.

The same tracking method

We monitored both groups for citations in GPT and Perplexity.

Measured over time

We kept running the same checks and counted citations and unique articles cited.

What happened

AI cited the articles with forecasts more often.

The gold line shows the 50 articles with Authority forecasts. The grey line shows the 50 plain articles. Each point looks back over the previous 24 hours, so you can see whether the difference holds over time rather than relying on one lucky result.

Updated every 15 minutes

How many times AI cited the articles

Every successful citation in GPT and Perplexity during the previous 24 hours. If one article is cited several times, every citation is counted.

Plain articles 95 Articles with forecasts 184

Updated every 15 minutes

How many different articles AI cited

The number of individual articles cited during the previous 24 hours. An article counts once even if AI cited it several times.

Plain articles 10 Articles with forecasts 19

The public graphs begin on 12 October, when the comparable reporting window starts. Last refreshed 13 October 2026, 19:40 London time.

Follow the study

Get the next result when the study updates.

The graphs are public. Your email is only used to send the weekly update.

What we added

A forecast gives an ordinary article something original to say.

We add a specific prediction that matters to the article's reader. The forecast has a live probability, a dated record of what changed, links to the evidence and commentary from real experts. It is useful information that AI cannot find on dozens of competing pages.

Authority.inc · real published example

Will the UK government bring back an Energy Price Guarantee before July 2027?

A live probability that moves when new market or policy evidence appears.

9.0%

chance today

1234567
Last updated 13 October 20267 sourced changes · reviewed evidence · expert commentary
6

New evidence

A new price-cap forecast made government support more likely.

The probability moved because a dated, attributable market event changed the outlook.

“The Budget must deliver additional targeted support for households most at risk this winter.”

Adam Scorer · National Energy Action

A clear prediction

A precise question with a current probability, not a vague trend summary.

Fresh, original data

The probability changes when new evidence changes the outlook.

A source trail

Every important change is dated and linked to the evidence behind it.

Expert commentary

Named specialists explain what the evidence means in the real world.

How Authority builds it

We build each forecast from data, evidence and expert judgement.

Our founding team comes from Bloomberg, the University of Cambridge and The Economist. Our forecasting methodology has been audited by Oxford AI. We are building forecasting models designed to be the most accurate in the world, then turning their output into plain-English predictions that people can understand and use.

01

Choose the right question

We find a future event that matters to the article and to the people searching for it.

02

Build the model

We use historical patterns, current signals and market data to estimate the probability.

03

Add human judgement

Relevant experts explain what the latest evidence means and where uncertainty remains.

04

Keep it current

We update the forecast as new evidence appears, so the article does not quietly become stale.

What this means for your website

Your old content can become useful again.

Most companies already have years of articles that answer valuable customer questions. The problem is that the content looks much like everyone else's and gets older every week. Authority adds information that is new, specific and worth citing.

We can turn a historical content backlog into a library of evergreen, high-authority pages that stay useful to readers and appear more often in AI search.

The article keeps what already works.

  • The same URL, search topic and existing backlinks
  • No website migration and no need to publish hundreds of new pages
  • A forecast matched to the exact question the article already answers
  • Fresh evidence and expert commentary added as the market changes
  • Ongoing measurement of whether AI search is citing the page

About the numbers

This study is still running.

We are tracking 100 informational articles in competitive B2B technology and operations topics. Fifty contain Authority forecasts and expert commentary. Fifty do not.

The results combine citations observed in GPT and Perplexity. The public graphs begin on 12 October 2026, when the comparable reporting window starts. The study is ongoing and the results will be updated weekly.

These are early, directional results rather than a completed or independently peer-reviewed academic study. We expect the difference may grow as Authority content stays current and ordinary content becomes older, but the purpose of continuing the study is to measure that rather than assume it.

AI softwareIndustrial automationEnterprise technologyOperations software

Make your existing content worth citing

Turn your best old articles into living sources for AI search.

We will show you which pages have the biggest opportunity and what an Authority forecast could add to each one.

Book a demo