Conversations model

The conversation model represents threaded discussions and communications — emails, chat threads, calls — that feed sentiment analysis, needle mover detection, and the account timeline. A conversation entry on the timeline shows the subject, sentiment, and a link back to the source.

Managed models from messaging connections

Many workspaces get a Conversation model automatically when you connect a supported messaging integration — for example Intercom, Slack, or Microsoft Teams.

1. Add the connection from Data connections.
2. Authorize and complete any required setup.
3. Under Configuration → Data models, confirm the Conversation model for that connection.
4. Refresh and validate on an account with recent conversations.

HubSpot: configuring email conversations

HubSpot email engagements are the most common source for conversation models. FunnelStory syncs them using the HubSpot v3 CRM API scoped to your deal companies, so only emails from accounts you manage are included.

If the HubSpot connection's Sync Emails option is on, FunnelStory already imports these emails as threaded conversations, so skip this model; configuring both imports every email twice.

Prerequisites

• A connected HubSpot data connection.
• The Accounts and Users models must be configured first. The email conversation query resolves account_id through the sender's email address matched to contacts_v2, which requires contacts to be synced.

Using Claude (recommended)

The fastest way to configure HubSpot email conversations is to ask Claude via the MCP server:

"Configure a HubSpot email conversation model"

Claude reads the data model guide, previews the query against your data, and creates the model in the right order. See Configure HubSpot models with Claude for a full walkthrough.

Manual configuration

Add a Conversation model with the HubSpot connection and the following query:

WITH engs AS (
  SELECT
    engagements.id,
    engagements.timestamp,
    engagements.data->'engagement'->>'bodyPreview' AS text,
    engagements.data->'metadata'->>'subject' AS subject,
    engagements.data->'metadata'->'from'->>'email' AS from_email,
    engagements.data->'engagement'->>'portalId' AS portal_id,
    contacts_v2.properties->>'associatedcompanyid' AS company_id
  FROM engagements
  JOIN contacts_v2
    ON LOWER(engagements.data->'metadata'->'from'->>'email') = LOWER(contacts_v2.email)
  WHERE engagements.type IN ('INCOMING_EMAIL', 'EMAIL')
    AND engagements.data->'engagement'->>'bodyPreview' IS NOT NULL
    AND engagements.data->'engagement'->>'bodyPreview' != ''
    AND engagements.data->'metadata'->'from'->>'email' IS NOT NULL
    AND contacts_v2.properties->>'associatedcompanyid' IS NOT NULL
    AND contacts_v2.properties->>'associatedcompanyid' != ''
  GROUP BY engagements.id, engagements.timestamp, engagements.data,
           contacts_v2.properties->>'associatedcompanyid'
)
SELECT
  engs.id AS engagement_id,
  engs.company_id,
  engs.timestamp,
  engs.subject,
  json_object('type', 'email', 'engagement_id', engs.id, 'source', 'hubspot') AS key,
  json_object('title', engs.subject) AS metadata,
  json_object('text', engs.text) AS encrypted_data,
  '<fs_analyze>' || engs.text AS sentiment,
  '<fs_analyze>' || engs.text AS text_analysis,
  '<fs_analyze>' || engs.text AS label_analysis
FROM engs
ORDER BY engs.timestamp DESC

Mappings:

Column	Property
company_id	account_id
engagement_id	conversation_id
timestamp	timestamp
timestamp	created_at
sentiment	sentiment
text_analysis	text_analysis
label_analysis	label_analysis
subject	email_subject
key	key
metadata	metadata
encrypted_data	encrypted_data

How account resolution works

The query joins email sender addresses to contacts_v2.email and uses the contact's associatedcompanyid to find the account. This means the sender must be a known HubSpot contact associated with a deal company. Emails from unknown contacts are excluded.

HubSpot notes

The HubSpot connection imports email engagements only, so a HubSpot notes model has nothing to read. Record account notes in FunnelStory instead (Adoption → Notes on the account), or bring them in from another source with a Note model.

Field reference

Model keys

Property	Type	Description
account_id	string	Account associated with the conversation.
conversation_id	string	Stable ID for the thread in the source system.

Default properties

Property	Type	Description
created_at	timestamp	When the conversation started or was opened.
sentiment	float	Optional aggregate sentiment score. Pass <fs_analyze> + the text to have FunnelStory compute it automatically.

Optional properties

Property	Type	Description
key	json	Structured key for deduplication: {"type": "email", "engagement_id": "...", "source": "hubspot"}.
metadata	json	Extra display data shown in the timeline tooltip: {"title": "...", "link": "..."}.
encrypted_data	json	Full text content stored encrypted: {"text": "..."}.
text_analysis	json	Structured NLP output. Pass <fs_analyze> + text to trigger automatic analysis.
label_analysis	json	Labels or topics extracted from content. Pass <fs_analyze> + text to trigger automatic labeling.
email_subject	string	Subject line for email-sourced conversations.
participants	json	List of participants for meeting or chat conversations.

Related

• Note model — for notes from other sources
• Support tickets
• Meetings
• Data models overview