⟨ BLOGPLAYBOOKSOUTBOUND

Replies triaged from my outreach tool → HubSpot + Slack

Every reply from your Instantly campaigns classified, logged on the right HubSpot contact, and the interested ones posted to sales Slack with suggested next steps.

PlaybookInstantlyHubSpotSlackRUN BY THE DEMAND GENERATION AGENT →

The problem this solves

Replies are the whole point of outbound, and they are handled worse than any other part of the motion. They arrive scattered across a sending tool's inbox, mixed together: a buyer asking for pricing sits between an out-of-office and an unsubscribe demand. Someone has to read each one, judge what it is, log the outcome on the CRM contact, and get the live ones to a rep - and during a busy campaign week, that someone is nobody in particular.

The cost lands in two places. Interested replies age: a prospect who answered within the hour deserves better than a rep discovering the reply on Thursday, and outbound's hardest-won asset - a stranger who engaged - depreciates by the day. Meanwhile the CRM stays blind: outcomes that never leave the sending tool mean HubSpot shows a contact as merely 'contacted' when the truth is 'referred us to their VP', and every downstream report inherits the blindness.

Unsubscribes are the quiet liability in the pile. A request to stop that goes unlogged because it sat in an unread tool inbox is a compliance failure waiting to repeat, since nothing stops the next campaign from emailing the same person again. Triage that runs on a reliable cadence closes that door on the same day the request arrives, every time.

How the mission runs

  1. Read the recent replies from Instantly. The agent collects the recent replies across your Instantly campaigns and reads each in full - the actual text, its tone, and what it asks for - rather than keyword-matching, because 'interesting timing, circle back in Q3' and 'not interested' deserve very different treatments.
  2. Classify every reply. Each reply is classified: interested, objection, referral, out-of-office, or unsubscribe. Referrals get the referred-to person captured; objections get the objection summarized; timing-based deferrals are distinguished from refusals, since one is a future task and the other is an ending to record respectfully.
  3. Log the outcome on the HubSpot contact. Each classification is written to the matching HubSpot contact - the outcome, the reply summary, and the date - behind the standard review gate. Unsubscribes are flagged for suppression so the request is honored everywhere, and the CRM finally reflects what actually happened in the campaign.
  4. Post the interested ones to #sales with next steps. Interested replies and warm referrals go to your #sales channel as claimable items: who replied, what they said, and a suggested next step - book the meeting, answer the pricing question, email the referred VP with context. The live conversations reach a human while they are still live.

The prompt

This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.

⟨ THE MISSION PROMPT · PASTE AND RUN ⟩

Read the recent replies from my Instantly campaigns, classify them - interested, objection, referral, unsubscribe - log each outcome on the contact in HubSpot, and post the interested ones to #sales with suggested next steps.

What comes back

A cleared reply queue with nothing lost: every recent reply classified and logged on its HubSpot contact, unsubscribes flagged for suppression, and the interested and referral replies posted to #sales with the reply text and a suggested next step each. Run on a daily cadence, reply handling becomes a same-day guarantee instead of a weekly excavation.

Make it yours

  • Tighten the cadence to twice daily during a launch week, when reply volume peaks and the speed of the first response matters most to conversion.
  • Route by territory: replies from enterprise accounts to the AE channel, the rest to the SDR channel, each with owner-appropriate next steps.
  • Add a weekly objection digest, since the pattern in why people say no is the cheapest positioning research your team will ever get.

Frequently asked questions

How accurate is the classification on ambiguous replies?

Replies are read in full rather than keyword-matched, which handles most ambiguity, and genuinely unclear ones are flagged as needing human judgment instead of being force-fitted into a category. The reply text travels with every classification, so a miscall is visible and correctable in seconds.

Does this write to HubSpot without review?

Outcome logging goes through the same staged-write gate as every CRM change: you see the batch of proposed updates and approve it. Teams that run the mission daily typically find the review takes a minute, because each item is a classification with its evidence attached.

What happens with unsubscribe requests?

They are treated as the highest-priority classification: flagged on the contact for suppression so future campaigns exclude them, and reported explicitly in the mission summary. An unsubscribe honored immediately is both a legal obligation and cheap goodwill; the triage makes sure none sits unread.

Go deeper

⟨ RUN IT INSTEAD OF READING IT ⟩

This mission runs minutes after signup.

Open a workspace, paste the prompt, and the Demand Generation Agent carries it end to end on your plan's monthly credits - evidence attached.

⟨ RELATED PLAYBOOKS ⟩