Enrichment

Smart Enrich — depth tiers, cost preview, run history, and publication refresh.

Smart Enrich builds rich member profiles automatically from public sources (LinkedIn, OpenAlex, Semantic Scholar, web research). You run it from the Enrichment section of a community's admin dashboard (in the Member profiles group), during the creation wizard, or as a bulk action on selected members in the People section.

Depth tiers

Every mode includes discovery, classification, and auto-fill; you choose how deep the research goes. Each level is a superset of the one before it.

TierBadgeWhat it adds
QuickLowest costLinkedIn merge when a match exists, with a basic web lookup. No in-depth research — ideal for large networks.
StandardRecommended+ Deeper, targeted web research to refine classification and fill in missing information, and a search for the LinkedIn profile of members who have none, or whose LinkedIn link an earlier run found leads to no profile or was not theirs.
DeepMost complete+ Specialized sources (publications, Doctolib, patents), applied automatically based on profile type.

Deep exposes Customize add-ons (expert) with two add-ons, each with a checkbox: Discover & sync publications (shows how many academic profiles are eligible) and Scrape Doctolib profiles (FR clinicians). A checked add-on only sets the ceiling — the worker still decides per profile whether each step applies (it won't chase publications for a non-researcher, nor search OpenAlex for the author profile of a member who answered, during onboarding, that they have never published academic work). Every Deep run also searches Google Patents, with no checkbox, for members who list patents and for those whose profile suggests they are inventors.

What you control before launching

The Smart Enrich form shows:

  • the scope — the selected members, or all eligible profiles. A profile that holds nothing but a name (and perhaps a personal email address) is left out and not billed, since searching on a name alone can attach a namesake's details to it. A profile needs both a first and a last name, and is enriched once it also has a headline, an organization, a bio or a LinkedIn profile URL, or, before it is claimed, a work email whose domain names an organization. A named profile with none of these is kept only when an earlier run began to undo its classification, on finding its research was about a namesake, and did not finish: the run keeps it to finish that undo, and can bill it. Started from the Enrichment or People section, the run's dialog says how many members it left out for having nothing but a name and, for a run on selected members, how many lack a first or last name; add their missing name, organization, headline or LinkedIn URL to include them;
  • a single Force update switch — off, the run picks every profile still missing at least one of its enrichment records, so a partially enriched profile is picked up again for the parts it lacks; on, it includes fully enriched profiles too and re-runs every step even if their data is up to date, which is useful after correcting a profile;
  • an Estimated time and an Estimated credits figure from the live rate card. The credit figure is a single number, not a range: each profile is billed the flat rate of the run's depth tier whatever optional steps fire for it.

If the estimated cost exceeds the community billing owner's balance, an amber warning appears: "This run may use up to N credits — more than this community's current balance…". When you may buy credits for that balance, it ends with Find my best deal, which opens Need more credits? on the Get credits tab of Settings → Billing in a new tab, set to this run's profiles and depth (see Billing). The warning does not disable the start button (Start enrichment, or Start Smart Enrichment in the creation wizard), which launches the run with no further confirmation.

Enrichment is billed post-hoc, at a flat per-tier rate: each enriched profile bills exactly one tier for the depth the run was started at. It is not metered per individual step — skipping an optional add-on does not lower the charge for that profile. See Billing for what is metered today.

API & AI tools

The switch that lets the community's admins start enrichment, add members and erase a member's data from the community through their own API keys and AI tools has its own section of the dashboard: API & AI tools, under Connections. Its card, Admins' API writes, holds the switch Admins can write through the API, which is on by default. Only whoever pays for the community can change it: the person who pays, or an owner or admin of the paying organization. See API keys.

Run history

Recent runs lists past and current runs. From a row you can View progress in detail, Resume a stalled run, or Cancel a running one; profiles already processed keep their data. Cancelling a run, or deleting the community, also stops a LinkedIn scrape the run has under way. While a run is active, a banner in the People section shows its progress.

When a run you started finishes or fails, you are told by push and by email, as your notification switches allow, if you still administer the community (see Notifications). A run started through the API or an AI tool, and the enrichment that follows an event's guest import, send no notice of their own.

Publication refresh

A separate Refresh member publications card fetches the latest OpenAlex papers for up to 200 members, in the background. Choose Normal refresh (members not refreshed in the last 24 hours) or Full refresh (every member with an OpenAlex author, however recently refreshed). It is available once every 24 hours; the card shows when the next refresh becomes available.

Data sovereignty

For claimed profiles (owned by a real member), enrichment results are stored as community-scoped suggestions and never overwrite the person's public profile directly. Fields sourced from protected origins (manual entry, a CSV import, a CRM sync, the profile interview) are never overwritten by automated enrichment, except on a profile no member has claimed yet, where LinkedIn's details are combined with them. When LinkedIn's headline or bio says something the profile's does not, enrichment adds it to the existing text and keeps every sentence of that text; when the two cannot be combined that way, the profile's text stays as it was. When combining them shows the LinkedIn profile to be someone else's, nothing from it is written. A link the profile already had is then held for an admin to approve or reject, unless it came from an Import LinkedIn Connections archive, whose sync counts as failed; a link enrichment found by searching is set aside, and any other match it found is tried. When none fits, one of them is still offered for an admin to approve or reject. The organization becomes the person's current employers, the ones their profile header shows, and a location that names only a country takes LinkedIn's city or region in that country; later LinkedIn runs keep it in that country. Each protected value replaced is saved, so Reset enrichment brings it back. A position from your import that LinkedIn shows as ended becomes a past position; a reset does not undo that, but importing the file again does. When the profile is then left with no current employer, an imported organization naming only employers LinkedIn shows as left, or employers only an earlier enrichment listed as current, is emptied; web research still searches with it, and a reset brings it back. On a profile no member has claimed, what web research or a Doctolib page says about the person is also added to a filled headline or bio in the same way, and a later LinkedIn sync in an enrichment run keeps it. A headline or bio with no record of where it came from is left as it is. Enrichment does not fill your community's custom fields.

Sur cette page