The Context Graph makes the latest connected customer information available in the scope where it belongs. When relevant information changes, Bigmind can review the affected Deal or Account and maintain the artifacts configured for it.
#Example: an expansion meeting
This example assumes the workspace has configured and activated MEDDICC, Deal Score, and Account Health Frameworks. Customers can create these kinds of assessments from templates or configure their own through the UI or MCP.
Imagine a meeting takes place for the Acme expansion Deal:
- The CRM association connects the Deal to the Acme Account. Activity attribution connects the meeting and its participants to those records.
- The transcript and meeting summary provide new evidence. A meeting Scorecard evaluates how that specific call was run.
- The Deal’s MEDDICC Framework can update Champion or Decision Process findings with evidence from the meeting. A Deal Score Framework can reconsider buyer engagement and next-step quality.
- The Account Health Framework considers the meeting as part of the wider Acme relationship, alongside other associated Deals, messages, notes, adoption evidence, and risks.
- If the meeting materially changes the Deal, Bigmind can update its rolling summary, reconcile its warning set, and suggest CRM changes for review.
The pipeline view can then show MEDDICC, Deal Score, and Account Health next to the CRM stage, amount, and close date. Selecting a result opens its findings, reasoning, and sources.
#What happens during a review
- Resolve the target: identify the Deal or Account affected by the event and its CRM relationships.
- Gather current context: use the records, attributed interactions, linked notes, and prior artifacts available for that target.
- Check what applies: resolve the active Frameworks whose target and applicability conditions match.
- Review the artifacts: evaluate the relevant Framework components and, for an open Deal, reconcile the summary, warnings, and CRM suggestions.
- Preserve traceability: keep supported findings tied to their sources and retain existing results when the available evidence does not justify a change.
See Artifacts for each maintained result and its lifecycle.
#What can trigger an automatic review
A relevant change can mark an associated Deal or Account for review. Examples include:
- A completed and associated meeting
- An attributed inbound or outbound customer communication
- A relevant CRM field change, such as stage, amount, close date, ownership, association, or lifecycle state
- A linked CRM note or related record changing
- A specific future checkpoint on an active Deal warning
Nearby changes for the same target are combined before review, so a burst of related updates does not create an AI run for every individual event. Importing or fully synchronizing CRM records makes the latest state available; it does not by itself run AI against every Deal or Account in the workspace.
#Automatic review and Refresh now
- Automatic review starts from a relevant change and checks the affected target and applicable artifacts.
- Refresh now performs a broader Deal review using the current CRM state and available evidence.
- Refresh account level reviews the complete associated Account relationship and its applicable Account Frameworks.
From a pipeline view, select one or more rows to start a refresh. Each target runs as a system chat that can be followed live, switched between, and revisited in chat history.
#Built-in reviews and Automations
Deal and Account artifact maintenance is built into Bigmind. Configurable workspace Automations remain separate and can perform additional work for your own process, such as sending a message or creating another output.
#Closed Deals
Bigmind uses imported CRM lifecycle flags instead of interpreting customer-specific stage labels. When a Deal is closed, ordinary automatic Deal maintenance stops, active warnings are resolved, and future warning checkpoints are cancelled. A user can still request a final refresh.
#Usage
Automatic and user-initiated Deal and Account reviews are tracked as AI usage and do not consume Action credits. See Actions and usage limits.
For definitions of records, relationships, evidence, configuration, and scope, see Concepts.