Provenance & Citations

In production

AI Design · Show your sources

Part of A Field Guide to Trust

Seen in production: the FinTech case → · the due-diligence case →

In the diligence platform the sign-off stayed locked until the evidence had actually been opened — and cycles still fell from three weeks to four days. That is what provenance buys: every claim grounded in a source the person can open, one click from the number to the document that produced it. When the stakes are real, people don’t act on a well-written guess; they act on something they can open and check.

The principle this page proves

A claim with its source beside it is evidence. The same claim without one is prose.

Provenance & Citations pattern A claim card linked to its three cited sources — one flagged as dissenting — and the original source document opened to the cited passage. Provenance & Citations Every claim traced to the source behind it. Claim Architecture risk: 72 3 sources Sources Q3 filing Eng. audit Founder call disagrees Source doc Cited passage

Provenance & Citations

Every claim traced to the source behind it.

  1. Claim

    Architecture risk: 72

    3 sources

  2. Sources

    Q3 filing · Eng. audit · Founder call

    Founder call disagrees

  3. Source doc

    Cited passage

Put the source next to the claim

The instinct is to keep the interface clean: show the answer, tuck the evidence behind a “details” link. But a source one click away is a source nobody opens. Put the primary citation inline, beside the number it justifies, so the evidence and the claim are read together.

This is the difference between a system that asks for faith and one that offers collateral. A risk score that reads “72% — based on these three filings” is one a person can begin to trust; “72%” alone is a verdict they’ll quietly override.

Try it — the pattern, live

The same live demo that runs in the book edition’s field guide. Synthetic data; nothing leaves the page.

Let them open the original, unedited

A citation that can’t be opened is decoration. The trail has to end in the actual artifact the model read — the filing, the ticket, the transcript — not a paraphrase of it. The moment a person can pull the source and see it themselves, the model stops being a black box and becomes a research assistant.

Heavy summarisation is the silent killer here. Compress the evidence too far and the trail goes cold; the person can’t tell whether the model read the document carefully or skimmed it. Keep the path from claim to primary source short and intact.

Count the sources, and surface the dissent

Say how many sources back a number. “Three sources” reads very differently from “one source,” and the person calibrates accordingly. A single-source claim should look more tentative than a well-corroborated one — the interface should carry that distinction.

Then surface disagreement honestly. If two sources support a conclusion and one contradicts it, show the outlier rather than averaging it away. The dissenting source is often where the real risk lives, and hiding it is how a confident-looking answer becomes an expensive one.

The strongest objection, kept in

“Citations clutter the interface and slow the reading.” In the diligence platform the sign-off stayed locked until the evidence had actually been opened — and cycles still fell from three weeks to four days. The clutter objection assumes reading speed is the bottleneck. It isn’t: re-verification is, and a claim that carries its source kills the re-verification loop entirely.

Ship it: the checklist.

  • Render the primary source inline, next to the claim — not behind a “details” affordance.
  • Make every citation openable: link to the real artifact, not a summary of it.
  • Show the source count, and let a single-source claim look more tentative than a corroborated one.
  • Surface the dissenting source instead of averaging it into the consensus.
  • Watch a real user: can they get from the number to the document in one move?

Trade-offs by Approach

ChoiceBest ForTrade-off
Inline citation chipsClaims a reader will challenge on sightCrowds dense layouts; needs disciplined sourcing
Drill-down drawerDeep evidence without leaving the flowOne click of friction; drawer must open fast
Source count + dissent flagSummary surfaces at a glanceA count can hide a weak source; flag the dissenter
Full document viewerAuditors, regulators, sign-off momentsHeaviest to build; reserve for the highest stakes

Where this pattern earned its keep

Get AI design patterns in your inbox

One pattern a month. The tradeoffs I paid for, plus the templates I use.

Where this breaks — do not use it here

Citation chips slow scanning — if every claim is cited with equal weight, nothing stands out and users stop clicking any of them. Reserve full provenance for claims someone must defend to a third party; summarise the rest. Provenance is a gate for decisions, not wallpaper for paragraphs.