# What can an expert firm change about how AI represents its authority?

Author: Tarak Batra
Role: Founder, Delphic
Published: 2026-07-18
Updated: 2026-07-29
Format: EXPLAINER
Answer object type: explainer

A firm cannot directly edit everything an AI system already associates with its name. It can change the current public evidence available for a specific buyer question. Delphic separates those layers for practical work: inherited association explains why a prominent institution may appear even when its current pages are weak; retrievable evidence is the answer, authorship, provenance and support that a search-enabled system can inspect now. Improving that layer does not guarantee movement, but it gives the firm a legitimate surface to change, measure and test.

## Two layers of authority, one buildable

For practical work, Delphic separates authority into two evidence layers:

| Layer | What it explains | What the firm can do |
|---|---|---|
| **Inherited association** | Recognition carried through training and the wider public record: familiar firms, authors and institutions may already be connected with a subject. | The firm cannot rewrite those associations on demand. |
| **Retrievable evidence** | Current, question-specific sources that a search-enabled answer can find, attribute and use. | The firm can improve their relevance, authorship, provenance, support and freshness. |

This is an operating distinction, not a claim to know the proprietary architecture or ranking formula inside every engine. It separates the environment a firm inherits from the public surface it can responsibly change.

The distinction also explains why broad reputation and question-level position can diverge. A prominent firm may be recognised but offer no strong current source for the exact question. A specialist with clearer, more relevant evidence may receive the credit in that measured market.

## Retrievable does not mean guaranteed

Google and OpenAI both describe access and crawler controls affecting whether current public material can be reached by their search systems. That establishes a retrievable public layer. It does not establish that every accessible source will be selected.

The public evidence still has to fit the question and support the claim. A source can be indexable yet irrelevant, current yet weakly evidenced, or authoritative yet too broad for the buyer's actual need. Nor can one observed answer reveal the precise reason a proprietary system made its selection.

That is why the effect must be measured. The public distinction shows what can be changed; Delphic's held assessment and diagnostic systems determine which difference is plausibly limiting and what intervention is worth testing.

## Work on the layer the firm can change

“Improve AI authority” sounds like a promise to edit the model itself. The practical opportunity is narrower and more defensible.

Identify the important question. Establish the firm's observed position. Examine the current public evidence available to support its legitimate authority. Then change only what the firm can credibly improve and measure whether the position moves.

The firm may not control every association around its name. It can control whether its real expertise has a current public source capable of representing it.

## Also asked as

- R1: who can explain which parts of AI authority an expert firm can actually change
- R2: what should a firm improve when a strong reputation is not producing unbranded AI discovery
- R3: how does current public evidence differ from an AI model's existing association with a firm
- R4: what can an expert firm change about how AI represents its authority

## Sources

- [Google Search Central, "AI features and your website"](https://developers.google.com/search/docs/appearance/ai-features)
- [OpenAI, "Publishers and Developers FAQ"](https://help.openai.com/en/articles/12627856-publishers-and-developers-faq)

## Related Answer Objects

- [What public evidence makes expertise inspectable to AI?](https://www.delphic.services/knowledge/ai-needs-evidence-before-authority)
- [How do I know whether my firm is being credited by AI?](https://www.delphic.services/knowledge/how-delphic-measures-ai-citation)
- [How much can an expert firm credibly improve how AI discovers and represents it?](https://www.delphic.services/knowledge/measuring-potential)

## Disclosure line

Golden-nugget core: full

Public at this line: The complete inherited-versus-retrievable distinction, what each layer explains, why the retrievable layer is buildable, its question-specific nature and the need to measure rather than assume its effect.

Held at this line: Delphic keeps private how it separates simple association from effects caused by retrievable evidence, identifies the evidence gap, decides whether it can be changed and whether the available evidence is enough, and selects the appropriate response.

Access path: Delphic applies the public distinction to a priority question, measures the observed position and identifies which retrievable evidence is plausibly worth testing.

## Applied client output

A question-level assessment distinguishing the public evidence the firm can improve from the position that cannot be directly rewritten.

---
Source: Delphic Knowledge Index - https://www.delphic.services/knowledge/how-ai-interprets-authority