They price change: the difference between what just arrived and what was already understood. Every public business is held in the market with a standing read that is constantly reshaped. The central estimate shifts, the range widens, the tails skew, and this new information is legible only when set against that standing read. Without it, the print on the desk is just paper.

Coverage is what every serious investor does on the names they truly own. It is the standing account they hold of the business: how it earns, the drivers that move the results, the model assumptions and market consensus, the debate, the open questions, the peer and value-chain read-throughs, and the signals that they and the market are watching.

The work is to maintain coverage by continuously mapping new information to the drivers, and refreshing priors against every new print, transcript, filing, expert call, sell-side deck, estimate, and peer datapoint. Coverage is the structured, evolving record of all of it; the prior against which this data flow is read. That is what a serious investor means when they say they have a stock covered.

Today, a coverage account lives only in the analyst's head. Every investor reconstructs it privately and holds it in working memory. The scale at which coverage can be implemented is bounded by the context one human can realistically carry. Even the best investors hold real coverage only on the names they truly own, forgoing it entirely or letting it go stale across the long tails.

Coverage is shareable in form; the laws of physics do not prohibit public facts of a business composing into one shared understanding layer. The inversion is to lift coverage out of one's head and into the layer its structural form has always implied. A layer which is maintained continuously at the breadth and depth no human team could sustain, refreshed against the inbound data flow the moment it lands.

What remains proprietary is judgement — the step between coverage and underwriting the position. This is shaped by whatever variant view, proprietary network, data, and understanding the fund brings on-top of the public account, in addition to considerations of the broader portfolio exposures and the fund's risk requirements.

Most AI for finance, however sophisticated or agentic, is built around the user's session. A prompt goes in → the LLM reasons → it retrieves some information → an answer comes out. There is no standing context or continuous reading, and each interaction starts without structured priors. This is summarisation, which explains why the results are shallow.

Investors do not re-understand the business every time they form a view, coverage is continuously maintained by updating against their pre-existing read. The alt-data wave proved this a decade ago. Once satellite imagery, card-panel and expert calls became broadly available, the edge in possessing them decayed. What separated the funds that continued to extract alpha from alt-data was the coverage it landed on. The utility of additional data inputs only scales as much as the layer that interprets them allows. The bottleneck was never summarisation.

Reuters' wire. Moody's ratings. Bloomberg's terminal. MSCI's indices. None of these did old work faster. Each was a shared layer that scaled a piece of the market's understanding, letting participants see, compare, and act on what had before then been private. Bloomberg consensus is only an aggregation of sell-side estimates, but once it is published every earnings print is measured against it, and a two-cent miss can take ten percent off the stock.

Each layer democratised a piece of the understanding required to participate, and once it exists in a shared form, no one reproduces it privately. Judgement is unshackled and able to move up the stack, onto the paths that are less defined and where differentiated interpretation actually matters. Every observation layer pushes the frontier of judgement further out.

The chatbot generation of finance AI does not do this. It lives inside the same paradigm — faster retrieval, faster summarisation, a slightly better surface — without standardising anything new. The next paradigm is not a better information surface but the shared understanding layer itself: Coverage at scale, across every name in the public market.

Wissen is the standing coverage layer of the public market. Continuous coverage of every public company at a depth and breadth no human team can hold and no current AI can sustain.

When new information arrives — a filing, an estimate revision, a value-chain read-through — Wissen reads it against the coverage already held. What was expected? What changed? Which driver does it touch? The output is interpreted change: the substrate judgement is exercised upon. Once it exists, the market does not operate without it. By democratising coverage, Wissen pushes the frontier of judgement up and onto the less-defined paths where the material impact of new information is still being decided, and where judgement determines the answer.

This judgement need not be human. As the AI economy builds across capital markets, every AI agent faces the same problem human analysts do where it cannot interpret information without a standing prior. Without a shared one, each reconstructs the world privately, with no traceable basis for what it produces.

Wissen is the substrate for both; a common infrastructure that human and machine judgement can read against providing continuous coverage with full source provenance, where every claim is deterministically traceable to the ground-truth source behind it.