Skip to content

Using the Viral Coach

Comparing a draft with your own past posts, and importing reference signals by hand. Nothing here predicts reach.

Updated

The Viral Coach compares a draft with posts already in your signal lake and shows the five most similar, together with whatever performance was actually recorded for each. It reports what your own signals say; it does not predict how a post will do.

It needs an embedding provider to work at all. Where none is configured the coaching form is not shown, and the page says so instead of offering a control that would fail.

BEFORE YOU START

  • An embedding provider configured on this deployment. This is a deployment setting an operator controls.
  • Signals in your lake. Published posts are captured automatically; anything else has to be imported by hand.
  • The "create" or "generate" permission to coach or to import; "view" is enough to read the lake.
  • Nothing else, and no credits.

STEPS

  1. Open Engage and choose "Viral coach", or go to /app/viral.
  2. Paste the draft into "Your draft" and choose "Coach this draft".
  3. Read the table: each row is a similar past post, how closely it matches as a percentage, and its recorded score.
  4. To add a reference by hand, use "Import a reference signal": paste the post text, and optionally its platform and source address.
  5. Choose "Import signal". Performance stays unknown unless you record figures with it - nothing is invented.
  6. Browse the "Signal lake" table to see everything captured or imported.

WHAT YOU SHOULD SEE

Up to five similar posts, ranked by how closely they match the draft rather than by score. A signal with no recorded figures shows "Performance unknown" rather than a zero.

WHAT THIS WILL NOT DO

  • It does not predict reach. It reports similarity against posts you already have.
  • It does not collect anything from outside the product. Nothing crawls other accounts, so an empty lake means nobody has imported anything, not that a collector found nothing.
  • It does not invent metrics for an imported reference.
  • The mobile app has no coaching at all. The mobile screen lists the signal lake and imports signals; comparing a draft is web-only.

IF IT DOES NOT WORK

  • "Coaching is not active yet." with "An embedding provider is not configured in this environment." is the standing state where none is configured. Signals are still being captured from every published post and will power coaching once it is enabled.
  • "Coaching is unavailable right now - the embedding provider did not respond." means your draft is untouched and your signals are still being captured. Try again shortly.
  • "No comparable posts yet - publish more and your lake will grow." means the lake has nothing embedded to compare against.
  • "Could not coach this draft. Please check your permissions and retry."
  • "Could not import that signal. Please check the fields and try again."
  • "The request did not complete, so it is not known whether this signal was imported. Retry once you have checked."

COMMON QUESTIONS

How is this different from the Content Coach?

The Content Coach asks what is worth changing about a draft, using nine named checks. The Viral Coach asks which of your own past posts the draft resembles and what those recorded. They need different providers and degrade differently: the Content Coach still runs five checks without a provider, whereas the Viral Coach switches off entirely.

Why is a suggestion's score blank?

Because that signal has no recorded metrics. Zero would say it performed badly; unknown says nothing was measured.

  • Using the Content Coach

    Nine named checks on a draft, five measured against a network's published policy and four judged by a model. Advice, never a gate.

  • Writing and approving content

    The content editor - saving versions, submitting for approval, approving, and why an author cannot approve their own work.