Agentforce AI Agent on Salesforce Leads, Contacts & Cases
Your WhatsApp message arrives from a number your Salesforce has never seen. Your Agentforce AI agent searches for a match, finds none, creates a new Lead, attaches the message, and replies, all before your reps see the notification.
Why record attachment is the hidden hard problem
Your inbound messaging tools usually usually stop at “receive the message”. Deciding which Salesforce record that message belongs to, new or existing Lead, Contact, Account, Case, is where most integrations fall over. Rules need to be explicit, matches need to be safe, and new records need the right fields populated.
Your Agentforce AI agent does this on every inbound SMS, text and WhatsApp message, in real time, without creating duplicates or mis-routing conversations.
Matching inbound messages to existing records
Your app checks the sender’s phone number against your Salesforce Contact, Lead, Account and custom objects. It applies your configured matching rules, exact match, normalised phone format, country code handling. If it finds one clean match, the message attaches to that record.
Your app follows your tiebreaker rules when it finds multiple matches: most recent activity, highest lifecycle stage, or a specific field preference you define in the admin setup.
Creating new records safely
Your sender doesn’t match an existing record, your Agentforce AI agent creates a new one. You configure which object to create, Lead is the most common, but Contact, Case or custom objects work too. The AI populates the phone number, timestamp, and source channel, and fires any standard Salesforce automation you’ve got on that object.
Your duplicate rules still run. Your validation rules still apply. The AI doesn’t bypass Salesforce, it uses Salesforce’s own creation logic, so the record you get is the same shape your team expects.
Case creation from inbound support messages
Your support-focused Agentforce AI agent, every inbound message can spawn a Salesforce Case. The AI classifies the message (billing, technical, feedback, cancellation) and creates the Case with the right record type, priority and queue. The original SMS, text or WhatsApp message attaches to the Case as the first activity.
Your support team opens their Case queue and sees real enquiries with real classifications, not a wall of undifferentiated messages to triage.
Starting conversations the other way around
Your record handling runs in reverse too. Your Agentforce AI agent can start a conversation on any record, send an SMS, text or WhatsApp from a Lead’s action bar, tie the message to that Lead, and let the AI handle the reply when it comes. Your team initiates outreach from the record they’re working; the AI continues it from the record the customer responds to.
Your inbound or outbound, the AI and the record stay linked. Your Salesforce timeline shows one continuous conversation. Your reports see one lifecycle, not two disconnected halves.