AI updates to customer and conversation data
Let agents save selected customer or conversation fields and require stored information before starting a task.
What it is
Update Data lets an agent save information to fields you explicitly select. A Workflow function can separately require certain stored fields before it starts a task. This helps the agent collect missing details without treating a statement in the conversation as proof that the details were saved.
Who uses it and why
Users with the Settings or Developer role decide which fields an agent may update and which details a task needs. Visitors provide information in the conversation, and the agent submits it through the configured function.
Where to find it
Create an Update Data or Workflow function under Settings › AI › Functions. You need Settings with AI allowed under Settings Permissions, or Developer.
An operator with the Settings role manages field definitions under Settings › General Settings › Data Fields. Tasks are configured under Settings › Bots › Virtual Assistant › Tasks, using the Bot Trainer role.
How it works
- You select Allowed data fields on an Update Data function. INSIDE derives the expected inputs from those field definitions, including available choices.
- The agent can see whether relevant fields are missing or populated. This status does not itself disclose their stored values.
- When appropriate to the visitor's request, the agent asks for missing information and submits it through Update Data.
- INSIDE rejects unknown fields, unsupported values, invalid choices and missing customer or conversation context. It validates the whole submission before saving any field.
- Accepted values use the existing field validation, masking and storage rules. The next AI request receives refreshed populated or missing status.
- Before a Workflow function starts its task, INSIDE checks the actual stored Required data fields. Missing fields prevent the handover, even if the agent claims to have collected them.
Already populated fields satisfy the prerequisite without asking the visitor again. A stored false or 0 can count as populated.
Configure it
With the Settings role, create the required field definitions first. For an Update Data function, select at least one Allowed data fields entry. Parameters are generated automatically, so you do not need to recreate each field manually.
For a Workflow function, select Task and add Required data fields. Leave the list empty when the task has no prerequisites. Assign both functions to the intended feature or skill when the agent needs to collect information before starting that task.
Use instructions that tell the agent when collection is appropriate. Missing fields alone should not cause an unrelated task to run.
Settings
| Setting | Where | Applies to | Default | Who can change it | What it does |
|---|---|---|---|---|---|
| Allowed data fields | Settings › AI › Functions | website | Not set; at least one must be selected | Settings role, Developer role | Limits an Update Data function to the fields it may submit. |
| Required data fields | Settings › AI › Functions | website | Not set; no prerequisites | Settings role, Developer role | Prevents a Workflow function from starting its task until these stored fields are populated. |
Fields
| Field | Where | Required | What it is for |
|---|---|---|---|
| Task | Settings › AI › Functions | No | Selects the task a Workflow function invokes after its prerequisites are satisfied. |
Permissions
| Permission | Lets you |
|---|---|
| Settings | Configure Data Fields, and AI functions when AI is allowed under Settings Permissions. |
| AI | Allows AI configuration within Settings Permissions. |
| Developer | Configure the AI functions and their selected fields. |
| Bot Trainer | Configure the tasks selected by Workflow functions. |
Limits and good to know
- Each function can select up to 64 allowed or required fields.
- Updateable fields must belong to the website and store customer or conversation information. Supported types include text, number, boolean, choices, date/time and rating.
- File uploads and fields without storage are not supported by Update Data.
- A submission needs at least one known, non-empty value. Omitting a field does not clear it.
- A validation error prevents all writes in that submission. If storage fails after saving begins, earlier accepted fields may remain saved.
- Saving a customer field does not verify the customer's identity or transfer the conversation.
- Workflow handover is supported for live and Voice Autopilot. Email Autopilot excludes workflow functions.
Depends on and works with
+1 more · this feature uses
- Captured data expiry and message sanitisationConfigured by
Saved values follow the field's expiry and sanitisation rules.
View all connections as a list
This feature uses
- Roles and permissionsConfigured by
Roles and AI permissions control configuration.
- AI feature settings and change historyConfigured by
The AI feature assigns functions and supplies collection instructions.
- AI functions and required inputsRequires
Update Data and Workflow must be available as configured functions.
- Reusable AI skills and instructionsEnhances
An active skill can supply the data-update or task function.
- Customer contact detailsEnhances
Approved customer fields can be saved to the relevant contact.
- Virtual Assistant task setupEnhances
Optional Workflow functions select an existing task.
- Virtual Assistant task data fieldsRequires
Existing field definitions control formats, choices, validation, masking and storage.
- Virtual Assistant tasks started by AutopilotRelated
Workflow handoff checks that required fields are stored before accepting control.
- Captured data expiry and message sanitisationConfigured by
Saved values follow the field's expiry and sanitisation rules.
Used by
- Virtual Assistant tasks started by AutopilotConfigured by
Required stored data fields must be populated before the task function accepts invocation.
- Autopilot handover and error handlingConfigured by
Required data must be populated before a configured task function can run.
- AI functions and required inputsEnhances
Update Data and workflow prerequisites control structured data collection.
FAQ and troubleshooting
Why is a field missing from Allowed data fields?
Check that it belongs to the website, has suitable customer or conversation storage, is not deleted and uses a supported field type. File uploads are excluded.
Why will the task not start after the visitor supplied the details?
The values must be successfully saved. Check that the agent has an Update Data function covering the missing required fields and that validation succeeded.
Does a false answer count as a missing value?
No. A stored false value or zero can satisfy a populated-field prerequisite. The task checks presence rather than whether the answer is positive.
Why were only some values saved?
Validation finishes before saving starts, but values are then saved in sequence. A storage failure can stop later saves while earlier values remain. Review the fields before retrying.