The first output is private
An expert firm's website is an incomplete account of what it knows. The real authority may sit across research, models, datasets, client deliverables, recurring methods and the judgement of specialists. Reading only the public site reproduces those omissions. Publishing the private corpus destroys the boundary the firm is trying to protect.
Private mapping separates three objects that are often confused:
| Layer | What it is | What happens to it |
|---|---|---|
| Source knowledge | The research, data, method and judgement the firm owns or sells. | Remains governed and private. |
| Question-market map | The private record of the buyer questions that knowledge can legitimately support. | Becomes a client decision system. |
| Public evidence | The small set of Answer Objects later approved for selected markets. | Crosses the public line only after a separate disclosure decision. |
Moving from source knowledge to the map publishes nothing. The map reveals the commercial shape of the firm's authority without turning that authority into free material.
The map begins where the website ends
The work begins with the actual knowledge estate. Existing research and data show what has been established. Methods and deliverables reveal recurring expertise. Expert interviews can surface judgement that has never been written down. Provenance, date, product relevance and sensitivity travel with the material so that an aspirational marketing claim is not mistaken for authority the firm genuinely holds.
External evidence then shows the market around that authority: the natural buyer questions, signs of demand, the public sources already supplying answers, the firms AI repeatedly credits and the client's measured position. Potential and disclosure analysis show where movement may be worth considering and where the knowledge cost is too high.
These evidence categories are public because the buyer should understand what the map represents. Delphic keeps private how knowledge is extracted and combined, how family boundaries are set, how different signals are made comparable and how the portfolio is ranked.
One map, several decisions
For every validated question market, the private map brings together:
- the authority and knowledge basis behind the answer;
- the buyer need and commercial relevance;
- the available demand and public-supply evidence;
- the firm's measured AI position and the current credited field;
- the attainable opportunity and disclosure boundary;
- the current portfolio decision and state of any approved public object.
This is one consolidated definition of the deliverable, not a content inventory and not a list of topics. It lets the firm decide what to measure, what to defend, what may justify public evidence and what should remain private.
The patterns can also reveal opportunities beyond publication: an underserved buyer market, a question the firm should build more expertise around, a product opportunity distributed across several engagements or a competitive movement that deserves attention. Those possibilities become part of the managed intelligence record rather than being discarded after an audit.
Privacy is an operating condition
“Under confidentiality” must describe how the work operates, not merely how it is sold:
- the engagement's confidentiality obligations apply before source material is received;
- source material and map outputs remain in a client-segregated workspace;
- access is limited to the named project team;
- the private corpus is not used as training material or transferred into another client's work;
- retention and deletion terms are agreed for the engagement;
- nothing is published without explicit content and disclosure approval.
The client controls every movement across the public boundary. Completeness applies to Delphic's private understanding of the firm's authority—not to how much of that authority the firm must reveal.