Trade show lead follow-up agent for healthcare and life sciences sales teams on Salesforce

What The Prospector does

Your team spends heavily on healthcare and life sciences trade shows, collects hundreds of badge scans and scribbled booth notes, then flies home to a full calendar. The scans land in a spreadsheet or a lead retrieval app export, get loaded into Salesforce days later, and sit untouched for weeks while the conversations go cold and competitors who were at the next booth follow up first. The Prospector works for the sales team and owns the window between the show floor and the first real conversation. It ingests every event lead, matches it to existing Salesforce accounts and owners, removes protected health information before any AI processing, enriches missing contact and firmographic data, scores each lead against your ICP, personas and booth intent signals, and drafts follow-up that references what actually happened at the show. It routes hot leads to reps the same day, runs approved sequences for the rest, and keeps working the event list until each lead books, disqualifies, or is handed off. The agent is built and operated as a managed service, so the event calendar, scoring rules and message library are kept current without your ops team carrying it.

Custom build

The team runs Salesforce, so the agent writes everything back to Salesforce leads, campaigns and tasks where reps and leadership already work, rather than introducing a new CRM. A custom Claude build fits better than an off-the-shelf SDR tool here because the hard parts are healthcare-specific: de-identifying booth notes before any model call, enforcing an approved claims library, and routing by ownership and compliance rules that vary by company. Claude can run in a BAA-eligible environment, n8n handles ingestion and orchestration with full control over what is logged, and Salesforce's own automation can still handle assignment rules if your admins prefer. Agentforce is a reasonable alternative if you are already licensed, but you would still need the PHI scrubbing and claims controls built around it.

What starts it

• A lead retrieval or badge scan export is uploaded to the shared event folder or pushed via the scanning app's API
• A Salesforce event campaign reaches its end date, prompting the agent to check that every campaign member has been processed
• An event lead in Salesforce passes the configured no-touch threshold without a logged activity from its owner
• A known event attendee visits the website after the show, detected by Warmly and matched to their campaign record
• A prospect replies to a post-show email, which the agent classifies and routes

What it reads

• Badge scan and lead retrieval exports, including booth notes, demo requests and session attendance
• Salesforce leads, contacts, accounts, open opportunities, owners and event campaigns
• FullEnrich waterfall enrichment and ZoomInfo for verified emails, titles, org hierarchy and facility or company firmographics
• Warmly website visit and buying signal data tied to event attendees
• An approved message and claims library maintained by marketing and medical, legal and regulatory reviewers
• The event calendar with booth themes, sessions your team hosted and products shown at each show

When a person takes over

• Hot leads (demo requested or high-fit buyer with strong booth intent) go to the territory rep the same day with a brief and draft, because a live conversation from the show deserves a human follow-up, not automation.
• Leads at accounts with an open opportunity or an existing customer relationship go to the account owner to decide whether and how to follow up, so the agent never cold-sequences someone mid-deal.
• Any booth note or reply containing possible patient information, an adverse event mention, or a clinical or off-label question goes to compliance or medical affairs, and the agent stops all automated contact with that lead until cleared.
• New message templates or any copy that adds a product claim go to the medical, legal and regulatory review group before they enter the approved library.

Rules it follows

• PHI never enters a prompt or log unless the model and logging stack are covered by a BAA. The design uses Claude through a BAA-eligible deployment (such as a cloud provider offering one), runs de-identification before every model call, and retains no raw booth notes in agent logs.
• Outreach may only use claims, outcomes and product descriptions from the approved library. Drafts are checked against the library before release, and any unmatched claim blocks the send.
• No automated contact with leads tied to open opportunities, existing customers, or anyone who opted out at the booth or since. Suppression lists are checked in Salesforce before every send.
• Respect institutional and HCP interaction policies. Messages to clinicians never offer meals, gifts or anything of value, and invitations to company events follow your compliance team's rules.
• Sending volume per mailbox is capped and all emails go from warmed rep domains, so a post-show surge does not damage deliverability.

How you know it works

• Median time from event close to first touch on every scanned lead
• Share of event leads with a logged first touch inside the agreed follow-up window
• Meetings booked per event, split by hot, warm and nurture tier
• Pipeline created and sourced to each event campaign in Salesforce
• Rep acceptance rate of agent drafts and briefs, tracked as a quality signal

From pilot to live

1. Pilot on a backlog and one upcoming showRun the agent on leads from a recent show that are still sitting in Salesforce, with every draft reviewed by reps and compliance. Use the same setup live on the next show with a small group of reps. Tune PHI detection, tiering rules and the claims library based on what reviewers flag.
2. Expand to the full event calendarConnect all lead retrieval sources, load the yearly event calendar, and turn on same-day hot lead routing and Chili Piper booking for every territory. Move warm-tier sends to batch approval once draft acceptance is consistently high.
3. Always-on managed operationExtend to webinars, regional events and conference sessions you sponsor. The operating team refreshes the claims library after each review cycle, retunes scoring after each show using closed-loop pipeline data, and reports untouched leads by owner to sales leadership monthly.