Research helps explain who a person or company is, why they may fit an audience, and what could make a conversation relevant. Distill combines professional background, company information, public sources, and supplied context to support audience decisions and message writing.
Research can be useful before any campaign exists. Enrichment covers matching identities, loading profile information, and finding or checking contact details.
What research can add#
A person’s research can include professional background, current and past roles, relevant company context, highlights, and public sources. Company research can include its business, products, market, funding context, and relevant public material where available.
These details help assess relevance: for example, whether a company has a public integration program, or which parts of a person’s experience relate to a partnership conversation. Profiles also contain ordinary factual fields; the Enrichment page explains those fields and identity matching.
Where the information comes from#
Distill uses stored professional-profile information, available company and public-web material, and supported data sources. Research can add highlights and source links that explain the background behind a profile.
Profile source records can include a title, URL, publication date where available, and a summary. The interface also exposes source details where available. Source links make the evidence behind a claim easier to inspect.
Customer-supplied information is another input. A list note, imported context, or a sender's own background can add information that public research would not provide. Where that information came from still matters: a customer note about a planned event is different from a public announcement that the person is attending.
Relationship information has its own coverage and meaning. A shared employer, a LinkedIn connection, and a possible introduction path are different kinds of evidence. None establishes friendship, endorsement, or willingness to introduce. Networks covers relationship data; Introductions covers possible routes through a connector.
Three kinds of research#
| Capability | What it does | Important boundary |
|---|---|---|
| Profile research | Develops or updates background research and highlights for a person or company. | Existing research can be reused when it meets the product's freshness rules. |
| Search traits | Assess a person or company against an audience question. Web-research traits can investigate public sources. | An unclear assessment is not proof of fit or non-fit. |
| List AI Questions | Answer a question for records in a list using profile information already stored in Distill. | Running a question does not by itself perform a fresh web search for every row. |
A search trait might ask whether a company publicly offers a partner program. A list question might summarize which parts of a person's existing background are relevant to that program.
List questions can produce an answer for each record, with reasoning and information about how confident the assessment is. They run as background work and can succeed for some records while failing for others. A completed batch does not imply that every row received an answer.
These outputs are AI assessments based on available evidence. They do not establish private buying plans, undisclosed budgets, or personal intent. Company search and People search explain how traits affect results.
Freshness and current employment#
Distill has separate updates for basic profiles, deeper research, and email information. They do not all refresh at the same moment.
When an assistant retrieves a profile, it receives existing information by default. Standard person and company research can currently reuse research from the last 30 days, provided that the research exists and has not been marked for updating. This is a reuse rule, not a promise to refresh every profile every 30 days or to verify every fact at that interval.
The profile returned to an assistant includes two separate status fields:
loadStatusdescribes whether the basic profile has loaded.researchStatusdescribes the research request.not_requestedmeans this read did not request research;pendingmeans the requested work is not yet ready;readymeans the research meets the reuse checks.
A loaded profile can therefore still have research pending or not requested. The detailed response does not provide one date showing when every field was last verified. A source's publication date describes the source, not when Distill last checked every claim.
Ready describes the relevant load or research state, not when each field was last checked. External sources can also lag behind a job change or leave an old position open.
Current-employer and email checks are separate from deeper research. Enrichment explains those checks and what a job change can affect.
A search result, a profile, and a prepared message can reflect information checked at different times.
How research supports personalization#
Research can give a message a concrete basis: a relevant role, a company initiative, a shared professional background, or an available relationship. Which information reaches an AI-written section depends on that section and the campaign's configuration.
For example, a fictional company profile might link to its public integration-partner page, and a person's current role might be Head of Partnerships. That can support a message about exploring an integration partnership. It does not support saying that the person requested an introduction, approved a budget, or personally recommended the sender.
Information can be available to message writing without appearing in the final copy. A list column is also not automatically a template variable, and an AI answer in a list is not automatically a campaign instruction. Campaigns and personalization explains message controls and context in full.
Working through a connected assistant#
A connected assistant can retrieve a detailed person or company profile and request supported profile research. Existing profile information is returned by default; a research request can reuse recent research or report work still pending.
The detailed response includes a selection of available information. By default it includes a limited number of relevant sources and recent work-history entries; supported options can request more. A basic batch profile response is a separate capability, not deep research on every person.
Profile loading and research report separate states. A partially returned profile can be pending; a failed load or unresolved lookup is not evidence that no such person exists. Enrichment explains profile loading and email-to-profile lookup. Using an assistant explains connection and tool availability.
Common questions#
Can Distill research someone before they are in a campaign?#
Yes. Profiles and research exist independently of campaigns. They can support audience research, meeting preparation, or understanding a person or company before outreach.
Is the information updated every time I open a profile?#
No. Opening a profile can reuse existing information. A research request follows separate freshness rules and can run in the background. Profile loading, research, and email checks are distinct.
Can Distill research something more specific than a job title?#
Yes. Search traits can assess a specific audience question, including supported web research. List AI Questions can answer questions from existing profile context. These are different capabilities.
Does asking an AI Question in a list search the web again?#
List AI Questions use profile information already stored in Distill. A web-research search trait is a separate way to investigate public sources. A list question can include reasoning and confidence, and individual rows can fail even when the batch finishes.
Can I see where a research detail came from?#
Profile sources can include a title, link, publication date, and summary where available. Web-research traits can return evidence and citations. The available sources help explain a result; a source’s publication date is different from when Distill last checked a field.
Can research make every message personal?#
It can supply relevant evidence, but quality depends on what is known and how the message is configured. Limited or unclear information may support a simpler message. Missing evidence should not be replaced with an invented connection or claim.