HubSpot Prospecting Agent Plays are instruction sets that tell the agent which companies to monitor, which buying signals matter, which personas to contact, and how to write the outreach. They are the core of the "Buying Signals and Contact Sourcing" beta that is replacing the older Selling Profiles setup. We build Plays for clients every week. Here is how the model works, what it costs to run, and the setup mistakes that quietly burn credits.
A Play is a reusable set of rules the agent follows. According to HubSpot's knowledge base, each Play is built across five tabs: Audience, Selling context, Outreach, Guardrails, and Automation. The agent watches the company segments you target for intent signals (funding rounds, product launches, research activity on topics you pick, website visitor intent), finds up to three contacts per company that match your target personas, and drafts personalized outreach for each one.
The model is in public beta. A Super Admin has to opt your portal into the beta named "Prospecting Agent Gets Smarter: Buying Signals and Contact Sourcing." Once it is on, Plays are available on Starter tiers and up across hubs, with HubSpot Credits required to run them. We covered the broader platform shift in our Agent Hub launch breakdown.
Older portals ran the Prospecting Agent on Selling Profiles, a single eight-screen wizard that defined one audience and one pitch. The Plays model replaces that with multiple Plays, each with its own audience, signals, personas, and outreach style. That is the real upgrade: you can run one Play per market instead of forcing one profile to cover everyone.
To check which version you have, go to Sales > Prospecting Agent. If you see a Create play button and tabs named Audience and Selling context, you are on the new model. If you are walking through Selling Profile screens, you are still on the old one, and HubSpot is migrating portals over time.
Audience: pick target company segments, choose the intent signals to monitor, add target personas, and set a daily new contact suggestions limit. The "Source contacts for me" switch controls whether the agent actively suggests new contacts.
Selling context: what you sell, your value proposition, and the pain points you solve. You can generate a first draft from your website URL, but edit it. Generic context produces generic emails.
Outreach: choose Sequence or Adaptive (compared below).
Guardrails: sender identity (contact owner with a fallback, or a single user), tone or brand voice, call-to-action options like a meeting link, and the send window.
Automation: "Review before sending" or "Send automatically." Start with review.
Want your Plays built from real research on your company? We build free personalized Breeze starter packs: signals, personas, selling context, and paste-ready agent setup, drafted from research on your business in about 4 hours. Get your free starter pack |
| Sequence | Adaptive | |
|---|---|---|
| How it works | Fixed multi-step AI sequence with editable steps, delays, and smart blocks | Agent decides each next email based on the prospect's actions, engagement, and signals |
| Control | High: you approve every step template | Frequency caps plus advanced instructions; the agent writes in the moment |
| Best for | Regulated industries, first rollouts, teams new to AI outreach | Warm segments and reactivation, where responsiveness beats consistency |
Build Plays inert, then flip them on. A Play can be published with "Source contacts for me" off and zero contacts enrolled. It costs nothing in that state. We build and review everything inert, then enable sourcing and enrollment only after the client signs off. Nothing burns credits while stakeholders are still reviewing copy.
Segments and personas are prerequisites. A Play targets company segments and personas that already exist in your portal. On a recent healthcare implementation, both were missing, and the build stalled until they were created. Check both before you open the Play editor.
The daily suggestion limit is your credit throttle. Credits are consumed when the agent generates personalized outreach for a lead. The daily new contact suggestions limit is the single most direct control over spend.
Turn off distress signals in sensitive industries. Outreach triggered by layoffs or bad news reads as ambulance-chasing. In healthcare and similar spaces we run Plays on research and growth signals only.
Pick one external contact source deliberately. The agent connects to Apollo, ZoomInfo, Surfe, or Seamless to source net-new contacts, but only one at a time. We compared options in our Seamless integration breakdown; if your ICP lives in Apollo or ZoomInfo, connect the one with the strongest coverage of your target market.
Every personalized outreach generation consumes HubSpot Credits, with additional consumption depending on Play settings. We published a full breakdown of agent credit costs and limits and a guide to Prospecting Agent pricing and access. If you have never turned the agent on, HubSpot offers a 14-day free access opt-in, which is enough time to validate one well-built Play against a warm segment.
Yes. A Super Admin opts your account into the "Prospecting Agent Gets Smarter: Buying Signals and Contact Sourcing" beta. Portals not opted in still run the older Selling Profiles version.
Up to three contacts matching your target personas per enrolled company, capped further by the daily new contact suggestions limit you set on the Play.
No. Only one external data source (Apollo, ZoomInfo, Surfe, or Seamless) can be connected at a time.
Not fully. Actions that consume credits do not run in sandboxes. Test in production with contact sourcing off and automation set to review before sending.
Yes, HubSpot offers 14 days of free access via an opt-in, available to Super Admins.
Want the whole engine built for you? Our AI SDR prospecting engine service designs the segments, personas, Plays, and guardrails, then hands your team a running system. See the AI SDR engine |
Questions about your specific portal? Talk to an expert.