Identity

Agents that remember customers and take on work for your team need to know who they are talking to. Settings > Identity is where you control that trust: which Slack, Microsoft Teams, and email accounts may hand work to Ask Dimedove on behalf of your teammates, and which system your agents treat as the source of truth for your contacts.

​

Team identities

A team identity links an outside account to someone on your team. Once linked, messages from that account are treated as coming from that teammate: they can hand work to Ask Dimedove, and their replies can approve what an agent sends, all recorded under that person’s name.

​

Confirmed automatically

Most accounts link themselves the first time a teammate writes in:

• Slack: a full member of the Slack workspace you connected, whose profile email matches a member of your Dimedove team
• Microsoft Teams: a user in the Microsoft 365 organization you connected, whose profile email matches a member of your Dimedove team
• Email: an authenticated email sent from the exact address a teammate signs in to Dimedove with

Guests, people from other organizations in shared channels, and bots are never confirmed automatically.

​

Unrecognized accounts

Any other account that writes in is held as an Unrecognized account at the top of the list, with Confirm and Reject actions. The person is told once that someone on your team needs to approve their account first. By email, this only happens for authenticated emails sent to your team’s own Ask Dimedove address; other emails from unknown senders get no reply. Dimedove holds up to five of their requests, and answers them as soon as the account is confirmed, provided that happens within seven days. Only an admin or owner can confirm an unrecognized account. To confirm one, pick the Team member it belongs to and read what it allows before you confirm. Pending confirmations also appear under Needs your attention on your Home page.

​

Adding and removing identities

Admins can link an account ahead of time with Add: choose the Team member, the Channel (Slack, Microsoft Teams, or email), and the account’s identifier (a Slack member ID, a Microsoft Entra object ID, or an email address). An account added this way is confirmed right away. To remove an account, choose Revoke from its ⋯ menu. A revoked account is never linked again automatically; admins can bring it back with Restore. A rejected account may ask again later. Each account can belong to only one teammate. Members see and manage their own accounts, while admins and owners manage everyone’s, including unrecognized accounts.

​

How agents recognize customers

Every conversation carries an identity level, based on how the person reached you. Your agents use it to decide how much of what they know they may use, from notes to details stored on the contact. Details such as addresses, postal codes, dates of birth, and policy, claim, licence, or passport numbers are treated as sensitive automatically. When something contradicts a person’s identity, for example someone else writing into a thread tied to one person, the conversation drops to Unknown and stays there until a teammate confirms who it is. See Confirming identity.

​

Source of truth

If your team keeps its customer records in a CRM or another record system, you can make it the source of truth your agents check before they reach out. It works with one connected record integration: HubSpot, Attio, Zoho CRM, Pipedrive, Close, Salesforce, Airtable, or Applied Epic.

1. Under Source of truth, click Configure
2. In Describe your records, choose the System (the integration) and the Object your agents should read
3. In How your team reads this object, explain what its fields mean to your team, in up to 1,000 characters. For example: “A deal in the Closed Lost stage means the customer asked not to be contacted again.” Only your team sees this description, and it is never shown to a customer
4. Save, then turn on the switch next to it (it is off by default)

Once it is on:

• Before reaching out for a goal, the agent reads the contact’s record to check whether it may contact them and what it should know. If the record cannot confirm the contact, the agent does not act.
• If the system cannot be read, for example during an outage, the agent waits and tries again rather than acting without it.
• In live conversations, agents can also draw on the record once the contact’s identity is Verified.