Turn every Gong call into competitive intel and Apollo lookalikes

After each new Gong call, an agent mines the transcript for competitors and objections, then queues 5 to 10 Apollo lookalike accounts in Slack.

Agentic Task
GongApolloSlackSalesLead EnrichmentResearch & MonitoringNotifications & Alerts

Build me an agent workflow that turns every new Gong call into a competitive intel brief plus a fresh Apollo lookalike account list, posted to Slack.

Trigger: Gong poll trigger on new_call. The agent runs once per new call as soon as Gong finishes processing it.

What the agent should do, in order:

1. Call Gong Get Call Transcripts for the new call ID. Read the full transcript and the call's account or deal context.

2. From the transcript, extract: any competitor names mentioned, objections or concerns the buyer raised, and recurring product themes or feature requests. If none are clearly present, say so explicitly in the final summary instead of inventing signal.

3. Use Apollo Organization Enrichment on the deal's account domain to capture industry, employee band, geography, technologies, and recent funding.

4. Run Apollo Organization Search to find 5 to 10 lookalike companies that match the same industry, similar employee band (within plus or minus one tier), and similar geography. Exclude the original account and any obvious customers or competitors.

5. For each candidate lookalike, run Apollo People Search to confirm at least one matching decision-maker title exists (for example VP Sales, Head of Revenue Operations, Director of Customer Success, or whatever titles fit our ICP). Drop accounts where no matching titles are found.

6. Run Apollo Organization Enrichment on the top 3 to 5 surviving lookalikes for one-line context (industry, size, a recent signal worth mentioning).

7. Use Slack Send a Message to post a single message to our sales channel. The message should include: a one-paragraph call recap, a bullet list of competitors mentioned and objections raised, and a numbered list of lookalike accounts with company name, domain, employee band, location, the matching decision-maker titles found, and a one-line reason this account is worth queueing for outbound.

Tone: tight and skimmable, written for an AE who has 30 seconds. Always include a link back to the Gong call. If the agent cannot complete a step (for example transcript not ready yet), it should fail loudly to Slack rather than post a partial brief.

Additional information

What does this prompt do?
  • Watches Gong for new calls and pulls the transcript as soon as it is ready.
  • Extracts competitor mentions, objections, and product themes that surfaced on the call.
  • Uses Apollo to enrich the deal account, then finds 5 to 10 lookalike companies with matching decision-maker titles.
  • Posts a Slack summary to your sales channel with the call recap, competitors heard, and lookalike accounts to queue for outbound.
What do I need to use this?
  • A Gong account connected with permission to read calls and transcripts.
  • An Apollo account with API access for Organization Enrichment, Organization Search, and People Search.
  • A Slack workspace and the channel ID where the post-call brief should land.
How can I customize it?
  • Swap the Slack channel for a per-rep DM, or split competitor mentions and lookalikes into two separate posts.
  • Tighten the lookalike filters to your real ICP (industry codes, employee bands, geography, technologies in use).
  • Adjust the decision-maker titles the agent searches for so the people surfaced match how your team actually prospects.

Frequently asked questions

When does the workflow run?
It runs on a Gong new call poll trigger. Each time Gong finishes processing a new call, the agent fires once for that call.
How does the agent find lookalike accounts?
It enriches the deal's account with Apollo Organization Enrichment to capture industry, employee band, and geography, then runs Apollo Organization Search with those filters and uses People Search to confirm matching decision-maker titles exist before recommending the company.
How many lookalikes does it surface per call?
Five to ten by default. You can raise or lower the cap inside the prompt, or have the agent only post when it finds at least N high-confidence matches.
Does the agent write to my CRM?
No. It only reads from Gong and Apollo and posts to Slack. If you want the lookalikes pushed into HubSpot or Salesforce as records, ask the agent to add that step.
What if a call has no clear competitor or objection signal?
The agent still posts a short recap and the lookalike list based on the account firmographics, and explicitly notes that no competitors or objections were detected so the rep is not misled.

Turn every sales call into a fresh prospect list.

Connect Gong, Apollo, and Slack once, and Geni runs this every time a new call lands.