Prospecting agent for small B2B SaaS teams on HubSpot that replaces Monday list building

What The Prospector does

Right now the reps at this 11 to 50 person SaaS company lose Monday mornings to manual list building: exporting from a data tool, cross-checking HubSpot for existing customers and open deals, guessing which website visitors matter, and pasting rows into sequences. Meanwhile trial users and accounts researching pricing go unnoticed until they go cold. The Prospector owns the top of the outbound and signal-led funnel for leadership. It holds the AI context (ICP, offer, proof points and common objections), builds weekly lists from firmographic fit plus buying signals, researches and enriches each account, drafts a personalized first touch and follow-up steps, applies suppression rules, and hands warm or high-fit accounts to the owning rep with a short brief. It does not send to anyone a rep or rule says to exclude, it does not negotiate or quote, and it escalates whenever an account looks like an existing customer, open opportunity or active product-led trial that sales should handle personally.

Built on HubSpot

The team already runs HubSpot, so ownership, deals, sequences, routing and reporting should stay there; Breeze prospecting and research features handle account summaries and draft emails inside the records reps already use. HubSpot falls short on net-new contact data, anonymous visitor identification and high-volume cold sending, so Apollo and FullEnrich supply data, Warmly reveals visitors and buying signals, and Instantly handles cold outreach from separate domains to protect the main domain. A custom AI agent orchestrated with n8n handles the weekly list assembly, scoring, product usage logic and suppression checks that span these tools, writing results back into HubSpot so leadership sees one funnel.

What starts it

• Scheduled weekly run every Sunday night so a prioritized prospect queue is waiting in HubSpot for each rep on Monday morning
• An identified company or person from a target account visits high-intent pages (pricing, integrations, comparison, security) above a set threshold
• A new trial or freemium signup from a company that matches the ICP, or a trial account crossing a usage milestone such as inviting teammates
• A tracked buying signal fires, such as a new hire in the buyer role, funding news or a job post mentioning the problem the product solves
• A rep marks an account as a target or a closed-lost deal reaches its re-engagement date

What it reads

• HubSpot CRM (companies, contacts, deals, owners, lifecycle stages, past email engagement and closed-lost reasons)
• Website visitor identification from Warmly or RB2B, synced to HubSpot company and contact records
• Product usage and trial data (signups, active users, feature adoption) pushed to HubSpot properties from the product database or analytics tool
• Contact and firmographic data from Apollo, with FullEnrich as a waterfall for missing emails and phone numbers
• A maintained AI context document covering ICP tiers, offer, proof points, objection handling, competitor positioning and banned claims

When a person takes over

• Any positive or ambiguous reply goes straight to the owning rep, because qualifying, booking and answering pricing or roadmap questions is a human conversation
• Accounts that are existing customers, open opportunities or active trials with real usage are routed to the account owner or founder for a personal touch instead of automated outreach, to protect expansion revenue and the product-led experience
• Enterprise or strategic target accounts above a set size get rep review of the research brief and first email before anything is sent
• Leadership approves changes to the ICP, offer, proof points and claims in the AI context, and the agent escalates any objection or question it has no approved answer for

Rules it follows

• Never cold-sequence an active trial user or an existing customer contact; product-led accounts get a rep-reviewed, helpful touch based on their actual usage, never a generic pitch that ignores they already use the product
• Only reference product usage data in outreach to people at that account who signed up, and never expose specific usage details to contacts outside the trial
• Use only approved proof points, customer types and claims from the AI context; no invented results, logos or features, and no pricing quotes in automated messages
• Respect unsubscribes, regional privacy rules and suppression lists across every tool, and send cold email only from secondary domains with warmup, verification and volume caps
• Every first touch must reference a real, sourced signal or research fact; if the agent cannot find one, the account goes to a review queue instead of receiving a generic email

How you know it works

• Rep hours spent on manual list building each week (target is near zero)
• Meetings booked per rep per week, split by outbound, website visitor and trial signal sources
• Positive reply rate and meeting rate per list type
• Pipeline created from agent-sourced accounts and trial-to-sales-assisted conversions
• Deliverability health: bounce rate, spam complaints and inbox placement on sending domains

From pilot to live

1. Pilot (weeks 1 to 4)Write the AI context with leadership, connect HubSpot, Warmly, Apollo and FullEnrich, and build suppression rules. Run the weekly list for one or two reps with every email approved by a human. Warm secondary domains in parallel before any cold volume.
2. Expand (weeks 5 to 8)Turn on visitor and trial signal routing for all reps, add Chili Piper to positive replies, and let low-risk mid-market follow-ups send automatically while enterprise accounts stay rep-reviewed. Launch the full-funnel dashboard and review it weekly with leadership.
3. Operate and improve (ongoing)Run the agent on a weekly cadence with monthly tuning of ICP scoring, messaging and signal thresholds based on reply and pipeline data. Maintain integrations, monitor deliverability and data quality, and update the AI context as the product, pricing and positioning change.