How to Put an AI Sales Team on Your Salesforce Leads

To put AI sales agents on your Salesforce leads, export a Leads report covering the records nobody has worked, then import that file into AI Sales Console, the AI Sales Brain: six specialized AI agents directed by one learning Brain. No connection gets switched on and nothing travels back on its own, so the exported file is the whole handover. A new account is live in 72 hours.

Start with the Leads object, not Contacts

Salesforce keeps unconverted Leads on a different object from Contacts, and that split does your first pass of qualifying. Contacts hang off an Account and often sit under an open Opportunity, so a handover list is almost always a Leads report. Standard fields carry the rest of the filtering:

• Lead Status, so Open - Not Contacted and Working - Contacted stop being one undifferentiated pile.
• Lead Owner, and whether a record sits in a queue instead of with a named person. Queue-owned leads are the clearest case of work nobody picked up.
• Last Activity, which separates leads a rep is progressing from leads logged once in March.
• Lead Source, when one channel's backlog should go over rather than every dormant record.
• Email Opt Out and Do Not Call, which are exclusions rather than filters. Filter either one out and the record never reaches the file.

From a live report to a frozen file

A Salesforce report is a query, not a list. It re-runs whenever somebody opens it, and its membership drifts as leads convert, change status, or get reassigned. The export is one snapshot of that query, and the snapshot is what the agents work, so treat the moment of export as the decision and keep the file.

Run the report and export it, or use Data Loader for a large or repeated extract. Bring the name, email, phone, employer and title columns across, and leave owner and score columns behind, since internal user names can be read as a person's name.

The importer takes .csv and .xlsx files, and any column it cannot place rides along as custom data. One checkbox gates it: the import button stays disabled until you attest to a lawful basis for emailing the list, and ticking it writes consent evidence against every lead created. Leads carrying a do-not-contact flag, or paused pending identity review, sit out until that resolves.

Who picks the leads up, and on which channel

The roster is Sage on research, Alex on voice calls, Mia on email, Zara on SMS, Nova on analytics, and Jade on coaching, with one Brain deciding what each does. A report row is thin: a name, a title, an employer. Sage finds the rest, and Mia holds the first email until that research is back. The Brain reads what comes back and gets sharper every week, and when a lead answers, the reply goes out in under 60 seconds, 24/7.

Plan the batch as email-led. Alex dials only where a recorded agreement to be called already exists, and exporting a report cannot create one. A clean Do Not Call column is a floor, not the permission.

Two systems now touch the same lead

The list left Salesforce. Your automation did not, and four things still reach for those records. Sales Engagement cadences are the loudest: an enrolled lead keeps getting your email alongside Mia's. Campaign membership is quieter and does not care the lead is worked elsewhere. Assignment rules refill the pile, routing new leads into the queue you drained. And an Agentforce SDR agent, if you run one, needs these records out of scope, or two agents write to one person. None of that stops a rep emailing one by hand, so say which list is spoken for.

Then settle the other half, because every touch the agents make is recorded in AI Sales Console, never on the Salesforce record:

That last row is the one that bites. A person who opted out here still reads as untouched in Salesforce, so your next export quietly pulls them back onto a list they already left. One named owner for those updates is what makes a second pass at an aging queue safe.

What it costs to work the list

Team is $1,999 a month for 250 leads worked, roster included, with no per-seat fees, and the ladder continues at $2,999 for 500 leads, $3,999 for 650, and $7,999 for 1,300. Anything past the allowance bills as an extra at $8 per lead, so scope each export to a month of capacity. Annual billing saves 17%, and eligible early-stage startups save 30%. The first 30 days run on your own pipeline, with no card, and there is a 30-day money-back guarantee on the first paid month. The salaried alternative, one person working that same report, costs $11,200-12,800 a month fully loaded (on-target earnings of $80,000 or more per The Bridge Group 2025, plus U.S. Bureau of Labor Statistics employer-cost factors, tools, and management).

When a handover is the wrong call

If your managers coach from the activity history on the record, a handover hides the outreach from them, because it reaches Salesforce only when somebody puts it there by hand. And if you already run Agentforce, Salesforce sells an agent layer that works these records in place, under the sharing rules and reporting you administer, which is the more honest fit when every touch has to stay on the Lead. A report returning a few dozen leads is a third case: one rep clears it in an afternoon. No software wins a complex, high-trust close. Humans do that. The AI is for the repetitive top-of-funnel volume that dies waiting for someone to work a list nobody was ever assigned.

The bottom line

Putting agents on Salesforce leads is an export, an import, and two standing decisions: whose automation stays off those records, and who moves Lead Status when somebody replies. For $1,999 a month, 250 leads aging in a queue get researched and emailed instead of forgotten, while the leads your report filters out for Email Opt Out or Do Not Call never reach the file. Every reply, and the work that earned it, sits in the live workspace, not in a Salesforce field somebody forgot to update.