# How can a Knowledge Index reveal where an expert firm should expand next?

Author: Tarak Batra
Role: Founder, Delphic
Published: 2026-07-23
Updated: 2026-07-30
Format: FRAMEWORK
Answer object type: framework

No single signal shows whether a buyer-question market is worth developing. The Knowledge Index joins four at the same granular unit: estimated demand, AI position and source ownership, movement in credible public supply, and evidence from measured treatments. Demand is itself an amalgamation, estimated primarily from traffic and live behaviour around each Answer Object and secondarily from sourced proxies. Together, the four signals show where interest exists, who currently receives the credit, how the competitive field is changing, what response the firm's knowledge is realising and which positions appear changeable.

## A market appears only when the signals meet

The question family is the common unit. Every signal must describe the same commercially meaningful buyer need before the evidence can be read together.

The architecture has four evidence lanes:

| Evidence lane | What it can establish | When it begins to read | What it cannot establish alone |
|---|---|---|---|
| **Question-family demand estimate** | An estimate of query volume and demand direction built primarily from granular Index traffic and live behaviour, then supplemented by sourced proxy signals. | A provisional read can begin from proxies. Its strongest first-party evidence accumulates once relevant Answer Objects and tracking are live. | A directly observed universal count of every query or the reason every individual visitor acted. |
| **AI position and source ownership** | Which sources appear, how stable the source field is and what role the firm receives. | From the first repeated study, then longitudinally. | How many buyers care or whether the position produces action. |
| **Public-supply movement** | Concrete changes in what the market publishes: new or materially revised sources, new authoritative entrants and knowledge becoming publicly available. | Initial source capture establishes the field; repeated observation establishes which publishing moves changed it. | Buyer demand or whether the firm can reproduce a rival's advantage. |
| **Treatment evidence** | Observed movement after a measured intervention, used to test and refine the earlier potential estimate. | After treatment and comparable remeasurement; it is not required for the initial market read. | General demand or the effect of a different intervention elsewhere. |

Demand needs one clarification. It is not a separate reading taken alongside traffic. It is an estimate built from an amalgamation of signals. Granular traffic and live behaviour in the Index are primary because every Answer Object maps response to a defined knowledge unit rather than a broad website topic. The record can connect a visit to its question family, identifiable referral, subsequent behaviour and content version. Sourced proxies add secondary context, particularly before enough live evidence has accumulated.

The client receives estimated query volume, direction and confidence for the family. Delphic keeps private which signals qualify, how activity is credited to a family, how different evidence is made comparable and combined, how gaps are handled and how confidence is set.

The joined record then supports two kinds of decision. It helps the firm decide what to do with markets already on its map: which require attention, defence, observation or restraint. It also helps identify adjacent questions, research needs, public evidence opportunities or knowledge products supported by more than one signal.

In practice, that can reveal:

- where demand, attainable movement and commercial value align;
- which mapped market now deserves attention and what changed;
- authoritative public supply entering before it becomes settled;
- protected knowledge becoming more widely available in the public market;
- the difference between a discovery problem, a low-response problem and an instrumentation problem;
- an adjacent question, content opportunity or product opportunity supported by several signals;
- which kinds of public-evidence change have produced measured movement before.

These are decisions the intelligence should support, not a public recipe for producing the recommendation.

## Each signal has a different limit

Demand remains an estimate because there is no directly observable universal counter for AI-question volume. Live behaviour is highly valuable positive evidence, but it is not a complete census: zero traffic can reflect low demand, weak citation, access or selection failure, incomplete mapping, faulty tracking, a short window or an answer completed inside the AI surface.

Public-supply movement is the publishing side of the market. A rival may release a new answer, materially strengthen an existing source, enter an adjacent question family or make previously scarce knowledge public. Those concrete moves show where the field is becoming crowded, where an incumbent is extending its authority and where an open market may be closing. They can also reveal that protected knowledge is becoming commoditised outside the firm, changing the value of keeping it gated.

Repeated observation is what connects the publishing event to a real change in public supply. A source appearing in one generated answer may be ordinary run variation; a new or materially revised source entering the measured field consistently is strategic evidence.

Potential estimates attainable headroom before an intervention exists. Treatment evidence, where available, adds observed movement under known conditions and helps calibrate what previously appeared attainable. It strengthens later expansion decisions but is not required for the initial assessment, and it does not imply that the same change will work in another family.

The signals also mature at different times. Proxies can seed demand while live Index behaviour accumulates. The initial repeated study establishes AI position, potential and the source field. Public-supply movement requires repeated capture. Treatment evidence begins only after an intervention and comparable remeasurement.

That timing belongs in the result. Where a signal has not begun to read, the record should say so. It should not fill the gap with an unsupported conclusion.

The public framework explains the signals and their limits. Delphic keeps private how unlike evidence is made comparable, how much influence each signal receives, how stale or incomplete evidence is handled, how confidence is set and how the joined record becomes a recommendation. A signal may direct investigation without deciding the treatment by itself.

Private client evidence remains segregated throughout. Wider market research can improve the external reference frame, but one client's protected material and results do not become another client's data.

## From answer position to market decision

AI-position work asks who is winning an answer and what role the firm receives. Question-market intelligence adds the larger questions: How large and active is this market? Is demand changing? Is credible supply entering? What response is the firm's knowledge realising? Does the position appear changeable? What should the firm defend, investigate or develop next?

The Knowledge Index can answer those questions because it does more than inventory publications. It preserves a longitudinal, question-level record of demand, supply, position, response and tested movement. That record becomes more useful as evidence accumulates; a later snapshot cannot perfectly reconstruct the conditions that came before it.

The result is not a forecast disguised as a dashboard. It is a managed view of the market at the unit where the commercial decision actually sits: the buyer question.

## Also asked as

- R1: who provides question market intelligence for expert firms
- R2: what evidence should guide where an expert firm publishes or develops products next
- R3: how can demand competitor movement and buyer response be read together
- R4: what can AI discovery data reveal about new markets for an expert firm

## Sources



## Related Answer Objects

- [What is a Knowledge Index?](https://www.delphic.services/knowledge/knowledge-index-for-expert-knowledge)
- [What is a question family (and what does one look like)?](https://www.delphic.services/knowledge/question-family)
- [How much can an expert firm credibly improve how AI discovers and represents it?](https://www.delphic.services/knowledge/measuring-potential)
- [Why does an expert firm need to keep managing how AI represents it?](https://www.delphic.services/knowledge/ai-citation-continuous-management)
- [How does Delphic estimate query volume for a question family?](https://www.delphic.services/knowledge/estimate-query-volume-for-question-family)
- [How does Delphic rank question-market opportunities?](https://www.delphic.services/knowledge/rank-question-market-opportunities)

## Disclosure line

Golden-nugget core: full

Public at this line: The complete market-intelligence architecture; estimated demand, AI position and source ownership, public-supply movement and treatment evidence; the primary role of granular Index behaviour and secondary role of proxies inside demand; the strategic decisions the joined record supports; and its confidence and privacy boundaries.

Held at this line: Delphic keeps private how live behaviour and proxy signals are selected, credited and combined into an estimate of query volume for each question family. This includes how missing data is handled, confidence is set, genuine publishing moves are separated from ordinary AI variation, opportunities are identified, unlike signals are read together, stale evidence is treated and priorities are chosen.

Access path: Delphic joins the available demand, position, supply and treatment evidence for the firm's mapped question families, then applies its held estimation and decision systems to return a confidence-labelled market record.

## Applied client output

A confidence-labelled, per-family intelligence record showing estimated query volume and demand direction, current AI ownership, supply movement, realised response, attainable headroom and evidence-led strategic opportunities.

---
Source: Delphic Knowledge Index - https://www.delphic.services/knowledge/question-market-intelligence