Changelog - June 10, 2026
Background Agents: AI that works after the conversation
Aimdoc now runs background agents — specialized AI tasks that operate separately from your Aimdoc agent. While your conversational agent handles visitors in real time, background agents analyze completed work and keep your knowledge base, visitor profiles, and CRM records up to date without slowing down the chat experience.
Background agent runs appear in a dedicated Agent (background) lane in the conversation timeline, so your team can see exactly what ran, when it completed, and what it produced.
- Runs automatically: Background agents trigger when a session becomes inactive, so insights are generated without manual effort.
- Separate from chat: These tasks do not interrupt or replace the conversational agent — they are a distinct, background actor.
- Transparent: Each run shows its status (queued, running, completed, skipped) and a summary of what it did.
- Built to grow: Background agents are a new platform capability. We will release additional agent types as new use cases emerge.
Knowledge extraction
After conversations end, Aimdoc reviews the transcript and extracts reusable knowledge your team can act on.
- Knowledge gaps: Surfaces business-relevant questions the AI could not answer, so you know where your knowledge base needs content.
- Human-answered Q&A: Captures Q&A pairs answered by a human rep during live chat, turning one-off answers into reusable knowledge (Business+ plans).
- Feeds your knowledge base: Extracted questions can be reviewed and added to your agent's knowledge, closing gaps over time.
Use case: A prospect asks about enterprise pricing and the AI says it doesn't have that information. Knowledge extraction flags the gap so your team can add a pricing page or FAQ — and the next visitor gets a complete answer.
Visitor summary
Background agents generate and maintain an AI-written summary of each visitor — pulling together conversations, browsing activity, events, and profile data into a single narrative your team can scan in seconds.
- Living profile summary: Updated as new conversations and activity occur, not a one-time snapshot.
- Rich context: Incorporates conversation history, visitor events, and browsing activity.
- Visible on visitor profiles and leads: The summary is stored on the visitor profile and attached to the lead when one exists.
- CRM sync: Map the Visitor Summary system field in Data Capture to push the summary into HubSpot, Salesforce, or Attio.
Use case: A rep opens a lead before a follow-up call and immediately sees a concise summary of every Aimdoc conversation, what pages the visitor viewed, and what they were trying to accomplish — without reading full transcripts.
Outcome classification
After a session ends, Aimdoc classifies the conversation against outcome categories you define — giving every conversation a structured result your team and CRM can act on.
- Custom outcomes in Data Capture: Define outcome categories (e.g., "Meeting booked", "Sent to support", "No clear outcome") in the Data Capture tab with a label and classification instructions for the AI.
- Automatic classification: The background agent selects the best-fit outcome when the session becomes inactive.
- Stored on conversations and profiles: The latest outcome is tracked on the conversation and visitor profile.
- CRM sync: Map Latest Aimdoc Outcome, Latest Aimdoc Outcome Label, and Latest Aimdoc Outcome Date system fields to sync outcomes into your CRM.
Use case: Your team defines outcomes like "Qualified — demo requested" and "Routed to support." After each chat, Aimdoc classifies the result so sales sees qualified leads in the CRM and support sees routed tickets — without manual tagging.
Relevant docs:
App Mode: deeper customization and a draggable widget
Since the April release introduced separate website and in-app configuration, App Mode has gained significantly more control in the agent builder — so signed-in users in your product get an experience tuned for onboarding and activation, not just a reskinned website widget.
Expanded in-app overrides
In the agent builder's App mode Configuration tab, you can now override engagement settings independently from your website agent. Each field inherits from the website by default and can be customized per override.
- Welcome message: A separate greeting for users inside your application.
- Conversation starters: App-specific starter prompts to guide in-app users.
- Links: Resource links shown at the top of the in-app chat.
- Input bar placeholder: Custom placeholder text for the collapsed widget input.
- Popup and auto-open: Independent popup message and auto-open timing for app mode.
- Prompt and appearance: Continue to override goals, colors, fonts, dark mode, and more on dedicated in-app tabs.
Draggable operator widget
The collapsed App Mode widget (operator icon) is now draggable. Users can reposition it anywhere on screen, and the position persists across sessions so the widget stays out of the way of your product UI.
- Drag to reposition: Click and drag the collapsed operator icon to any screen position.
- Persists across visits: The chosen position is remembered for returning users.
- Stays on-screen: The widget automatically adjusts if the viewport resizes.
Relevant docs:
Conversion funnel analytics
The analytics dashboard now includes a conversion funnel view that visualizes how visitors move through key milestones.
- Configurable steps: Track Visitors → Session Started → Lead Created → Calendar Shown → Meeting Booked or Escalated.
- Toggle steps: Show or hide intermediate steps to focus on the metrics that matter.
- Drop-off visibility: See conversion rates between each stage to identify where visitors fall off.
CRM and identity sync improvements
Background agent outputs and visitor data now flow more reliably into your CRM.
- Outcome and summary sync: Visitor summaries and conversation outcomes sync to HubSpot, Salesforce, and Attio via Data Capture field mapping.
- Lead identity mirroring: Captured lead identity data is mirrored to visitor profiles for a consistent view across the product.
- Inactivity-triggered sync: CRM fields update when conversations become inactive, alongside background agent runs.
Relevant docs: