Skip to content

Capability 04 — Authority & Digital PR

Answers are assembled from sources. The question is whether you are one of them.

When we trace why a competitor keeps getting recommended, the evidence chain almost never runs through their homepage. It runs through a review platform, a trade publication, a practitioner thread or a piece of research they published two years ago.

Why this is load-bearing

You can only say so much about yourself.

A claim made on your own website is a claim. The same claim confirmed by three independent sources is something a system can repeat with confidence. That asymmetry is why authority work is not a nice-to-have in AI search — it is often the binding constraint.

The diagnostic that starts it

For every tracked query where a competitor is recommended and you are not, we collect the sources the answer drew on and read what they say. That produces a specific, evidenced list of the places you are missing from — not a generic recommendation to “build more links”.

Unlinked mentions count too

A mention with no hyperlink does nothing for classical link metrics and still corroborates the entity. We track them separately from backlinks for exactly that reason.

Process

How an opportunity becomes a placement.

  1. 01

    Evidence-chain analysis

    automated research
    Collect the sources cited in answers across the tracked query set. Identify which source types and specific properties are doing the work in your category.
  2. 02

    Gap and opportunity mapping

    automated research
    Where competitors appear and you do not. Scored by relevance to tracked queries, plausibility of a win and effort.
  3. 03

    Angle and asset development

    human led
    What we actually have to say that a publication would want. Usually built on expert interviews or proprietary data.
  4. 04

    Human approval

    mandatory gate
    Nothing goes outside without a named person approving it. In fintech that usually includes compliance, and we plan the timeline around it.
  5. 05

    Outreach and follow-up

    human sent
    Targeted, relevant, capped in volume. Research and personalisation are assisted; the send is a human decision.
  6. 06

    Tracking

    continuous
    Placement, referring domain, brand mention, and whether the source starts appearing in the evidence chain behind tracked queries.

Surfaces

Where the work happens.

Prioritised per client against the evidence chain, not against a generic list of domains with high authority scores.

Surface
Industry publicationsFrequently cited as sources in category answers, and read by the same buyers.
Review platformsCategory pages on review sites appear constantly in the evidence chain behind “best X” answers.
Practitioner communitiesThreads where people ask peers what to use get retrieved and quoted.
Original researchData nobody else has is the most citable asset a company can produce.
Podcasts and newslettersTranscripts and archives are indexed; the reach compounds in the niche.
Partners and integrationsPartner directories and integration pages corroborate what you connect to and who works with you.
Analyst and directory listingsStructured category placement that reinforces what category you belong in.
Customer storiesNamed, specific outcomes are the evidence a recommendation can rest on.
Hard limits

What we will not do, at any budget.

These are not terms-of-service concerns dressed up as principles. Manufactured authority is detectable, it decays, and when it is discovered the damage lands on the client, not the agency that built it.

  • No fake accounts, on Reddit or anywhere else.
  • No account farms, sockpuppets or persona networks.
  • No fake, incentivised or written-for-you reviews.
  • No automated or undisclosed community participation.
  • No artificial mentions whose only purpose is to manipulate a search result.
  • No paid placements presented as editorial.

Automation boundary

What the agents do, and where a human takes over.

This is a formal policy, not a preference. It is written into the agent specifications in this repository and enforced as an approval gate.

Automated

Research. Opportunity discovery. Finding relevant discussions. Journalist and publication research. Outreach research and personalisation inputs. Content briefs. Draft responses and draft pitches. Tracking and reporting.

Human, always

Every external send. Every publication. Every community post. Every statement made in your name. Approval is recorded with the approver and the timestamp, so there is an audit trail if anyone ever asks.

Community participation

Real identified people, with real expertise, disclosing their affiliation, adding something useful. If nobody at your company can credibly do that in a given thread, the correct action is to not participate.

See where you stand in AI search.

An AI Search Audit tells you how often AI systems name your company, who they name instead, and what is causing the gap. Every figure comes with the method behind it.