The AI Prospecting Agent searches for new accounts in your ICP, enriches them via Apollo.io or Clearbit, places them in HubSpot and prepares the first outreach. Outbound automation that gives your SDR a filled queue at the start of the day, not an empty screen and a long coffee.
The Prospecting Agent is not an Apollo licence and not a LinkedIn scraper. It is a work unit that continuously searches your ICP, enriches with firmographics and intent, ranks, and then lands in HubSpot with a first-contact action ready for your SDR to review.
Most outbound teams spend 60 to 70 per cent of their time on account sourcing and data work rather than having conversations. That is exactly the part an agent reliably takes over, provided your ICP definition is sharp and your outreach voice is clear. The Prospecting Agent does not claim to be better than your SDR, it does make the SDR's day different.
Four things you want to know before you go live with an agent: which data it uses, which tools, how it lands in HubSpot and what the KPIs are. Click each topic for the details.
Data and sources
The agent runs on a combination of Apollo or Clearbit for firmographics, technographics and contact data. Additionally, we pull in public sources for signals such as funding rounds, new hires or vacancies that indicate your ICP. Everything is grounded on your ICP definition, previously won deals and your positioning. This way the agent does not source accounts that appear on some list, it sources accounts that belong on your list.
Stack and integrations
Default stack: HubSpot Sales Hub as system of record, Apollo or Clearbit as data source, HubSpot Breeze Prospecting Agent or a custom OpenAI or Claude flow for scoring and personalisation. For outbound sequencing we use HubSpot sequences, not a separate outreach tool, so everything stays in one system. The agent delivers results via the HubSpot API and logs every action as an activity on the Company record.
HubSpot landing
New accounts are created as Companies with the correct lifecycle stage, owner and custom properties for agent score and source. Every Contact and Company receives an audit-trail note. The first outreach lands as a task for the SDR, not as an automatically sent email. This keeps human review at the front and the SDR remains ultimately responsible for the tone.
KPIs and review
Upfront we choose three to five KPIs per engagement. The defaults are: reply rate on first outreach, quality score of sourced accounts (manual spot-check per week), SDR time saved per week and pipeline conversion from agent source to qualified opportunity. We build the dashboard together in HubSpot. Monthly review with the sales lead to tune grounding and scoring.
A B2B SaaS organisation had two SDRs and a sales lead who were spending half their time in Apollo and LinkedIn. No time to have conversations, too much time manually qualifying accounts. We built a Prospecting Agent on their ICP definition, with Apollo for enrichment and HubSpot as system of record.
After six weeks every SDR had a queue of 15 qualified accounts with a first-touch task every morning. The SDRs reviewed for 10 to 15 minutes and then went straight into outreach. Reply rate on first outreach rose from 8 to 14 per cent because personalisation became sharper.
On a Prospecting Agent engagement you work with Carel as agent architect and Yoni as sales and GTM sparring partner. Yoni helps you sharpen the ICP definition, Carel builds and runs the agent.
Honest answers to the questions we hear before every engagement.
In HubSpot the agent runs on your CRM data and buying signals. It sources accounts that score against your ICP definition, enriches them with firmographic and intent data, and prepares a draft outreach per account as a task for your SDR. For every account you see a brief rationale with the source data, so there is no black box. The agent works within your existing lifecycle and deal stages, so a qualified account lands in the right place in the pipeline immediately. We build this natively because we are a HubSpot Platinum Partner.
No. The agent completes the top-of-funnel work: sourcing, enrichment, scoring and first-outreach preparation. Your SDR does what an SDR should do: review, start the conversation and hand off to Account Executive. What the agent wins back is time. What the SDR gets back from it is reply rate.
Three things: a sharp ICP definition (or willingness to sharpen it with us), a working HubSpot tenant with lifecycle stages and deal stages, and an SDR who reviews during the first month. We also need your Apollo or Clearbit licence, but if it is not in place yet we sort that out in the first week.
Not by default. The first outreach lands as a task for your SDR with a draft message ready. We choose reply rate over volume, and a human at the front keeps the tone sharp. For clients who later want to scale to semi-automatic outreach that is possible, but only after the KPIs have been stable for two months and the SDR trusts the output without checking.
Two safeguards. First: the agent scores every account against your ICP definition and only places accounts above a threshold. Second: your SDR sees a brief rationale with the source data for every sourced account. Spot-check per week, and if quality drops we tune the grounding. No black box, but an audit trail per account.
The agent works best in HubSpot because we are a Platinum Partner and the integrations work natively. A similar build is possible for Salesforce, but it is a longer engagement. For Pipedrive or a niche CRM it starts with a sparring call: get in touch if your situation differs.
Then we do not start with this agent. A Prospecting Agent on a HubSpot without working lifecycle stages or without an ICP definition will cause chaos. We start with the Maturity Scan or an AI Readiness Scan to see where you stand.
Book a strategy call. We review your current outbound work, ICP definition and HubSpot setup, and give an honest scope. Not a sales call.