Setup

Indexing your whole site is the most common way to make an assistant worse.

The instinct on setup is to point the crawler at the domain and let it read everything. It is the single decision that most reliably degrades answer quality, and it is almost always made in the first ten minutes.

Before and after on a setup siteThe same visitor question, handled by the current site and then by an assistant grounded in that site's own content.BEFOREPoint the crawler at thedomain, index everythingAssistant confidently quotesa 2023 blog post as currentpolicyTeam assumes the model isweak and upgrades itNothing improves, because thewrong passage still winsAFTERIndex the pages you wouldsend a customer to, andnothing elseEvery source has an ownerand a reasonRetrieval competes betweengood candidates onlyRefusals rise, which ishonest, and answer qualityrises with them
  • Private launchOnboarding selected teams now
  • Built by CreoglyphSeven client engagements behind the product

Why more sources make answers worse

Retrieval ranks passages by similarity to the question. Every additional source adds candidates. If those candidates are outdated blog posts, superseded policies or marketing pages that describe an aspirational version of the product, they compete with the correct passage and sometimes win.

When a stale passage wins, the model does exactly what it should: it reads the supplied text and answers from it, confidently, because from its position that text was the relevant context. The failure happened before the model was involved, and no amount of model quality fixes it.

This is why teams who upgrade their model to fix accuracy usually see no improvement. The problem was never phrasing. It was that retrieval returned the wrong passage from a source that should not have been indexed.

What changes

Today

  1. Point the crawler at the domain, index everything
  2. Assistant confidently quotes a 2023 blog post as current policy
  3. Team assumes the model is weak and upgrades it
  4. Nothing improves, because the wrong passage still wins
  5. Trust in the assistant drops and it gets removed

With an assistant

  1. Index the pages you would send a customer to, and nothing else
  2. Every source has an owner and a reason
  3. Retrieval competes between good candidates only
  4. Refusals rise, which is honest, and answer quality rises with them
  5. Gaps are logged so the source list grows deliberately

A source list you can defend

Start with the sitemap

It is a list you already maintain of pages you consider canonical. It is the best starting point and it is usually already too broad.

Cut to pages you would link a customer to

If you would not paste the URL into a reply, it should not be a source. This one rule removes most of the problem.

Remove anything dated

Blog archives, changelogs older than the current version, superseded policies. If it was true and no longer is, it is worse than absent.

Add the files that are not on the site

Handbooks, specification sheets, price lists. Often the highest value sources and usually not crawlable.

Give each source an owner

Not a tooling feature, a working practice. A source with nobody responsible for it is a stale passage waiting to happen.

Reindex on publish, not on a schedule

Scheduled reindexing is simple and always slightly wrong. Event triggered is correct and requires your publishing system to emit an event.

The three ways a conversation leaves the flowMost conversations end in a grounded answer. The exits are a logged refusal, an escalation to a person, or a capture before continuing.IN THE FLOWQuestionRetrieveGrounded answermost conversationsEXITSRefuse, log gapEscalateCapture, then continue

A pruning pass you can run in an afternoon

Export the list of URLs you are indexing. For each one, answer three questions in writing.

Would I send a customer this link today

If no, remove it. This alone typically removes a third of an over broad index, and it removes the third most likely to produce a confidently wrong answer.

Is anything here contradicted by another page

If yes, fix the contradiction before indexing either. An assistant with two conflicting sources will pick one and will not tell you it had a choice.

Who notices if this goes out of date

If the answer is nobody, either assign someone or remove it. Ownership is the only durable defence against staleness.

Run the same pass quarterly. An index is not a setup step, it is a maintained artefact, and the failure mode is always the same: it was right on the day it was built.

Where the data goes

The assistant answers only from sources you nominate. Content behind a login, inside an image, or rendered only by client side script is not reachable by the crawler and will silently be absent rather than fail loudly. Security and data boundaries.

Questions

Fewer than you think. Ten well chosen pages outperform two hundred indiscriminate ones, and starting small makes it obvious which gaps are real.

Usually not the archive. A blog post is a snapshot of an opinion on a date, and an assistant will quote it as current fact. Index the handful that are genuinely evergreen reference material.

Not reachable. The crawler sees what an anonymous visitor sees, so gated documentation has to be uploaded as files if you want it indexed at all.

Uploaded files are supported within the storage limit on your plan. A text based PDF extracts well, a scanned one does not, because there is no text to extract.

On publish, if your system can trigger it. Otherwise weekly, and immediately after any change to pricing, policy or hours. Those three are where a stale answer costs the most.

Read the citations on wrong answers. A source that keeps appearing in answers you would not have given is a source to remove, and that is visible in the transcript log.

Yes, by plan: one on Free, five on Starter, ten on Pro, unlimited on Scale. See pricing. The limit is a feature on the lower tiers more often than it is a constraint.

Signal

Is this your situation?

Access is by waitlist, demo request or public-content pilot. Describe the question your visitors keep asking and we will tell you honestly whether it fits.