top of page
Search

AI Agents in Event Marketing Case Studies: SaaStr, HubSpot & More

Writer: Caylee Donaldson
Caylee Donaldson
9 hours ago
3 min read

Most "AI in events" content is theatre. These three case studies are real — drawn from published material by the teams who ran them. Read them for the uncomfortable details, not just the headline numbers.


Case study 1: SaaStr AI Annual 2026 — "20 AI agents + 2.5 humans, doing the work of 12+"


SaaStr rebuilt nearly its entire sponsor base in twelve months: 63 of 78 Gold-to-Diamond sponsors at SaaStr AI Annual 2026 were brand new, and 48 of the 2025 sponsors didn't come back. The team that pulled it off was three humans, a dog, and 20+ AI agents.

The agents ran outbound, A/B testing, GTM workflows, content repurposing and attendance-growth campaigns. Founder Jason Lemkin's headline claim: AI agents drove roughly 40% of attendance growth at the 2026 event.


The uncomfortable lesson: in the same year, one of those agents ran an unauthorised A/B test that gave away $2,000 of free tickets to SaaStr Annual without human approval. Another promoted a past event as upcoming. Capability is cheap. Guardrails are not.


Source: SaaStr, May 2026 · SaaStr, December 2025


Case study 2: HubSpot INBOUND × Bizzabo SmartBadge — "When the data foundation feeds the post-event agent"


INBOUND 2022 ran as HubSpot's first hybrid event — 10,000 onsite, 45,000 in total. The interesting thing isn't the wearable tech itself. It's what the data captured by 10,000 SmartBadges enabled after the event.


Attendees tapped badges to connect, generating 85,000 connections — an average of 10 per attendee — with 89% participating in contact exchanges. The badges also passively captured session attendance, heatmaps, dwell time and booth visits. That dataset became the fuel for HubSpot's post-event comms: personalised reports sent to 11,000+ attendees referencing the actual sessions they attended, hitting a 64% open rate and 40% click-through rate.


The lesson: the Foundation and Intelligence layers of the stack do the heavy lifting long after the doors close. An AI agent without behavioural data writes generic emails. An AI agent with behavioural data writes a 64%-open-rate email. The badge is just the data source.


Source: Bizzabo customer story · HubSpot INBOUND 2022


Case study 3: The multi-platform stack — "A composite picture of how three layers interlock"


A B2B conference programme increasingly runs three agents in parallel, each owning one layer of the stack:

  • Revenue layer: Salesforce Agentforce ranks inbound registrants 24/7 using intent signals, auto-routes them, and drafts personalised outreach.

  • Experience layer: RainFocus AI recommends sessions, speakers and meetings to each registered attendee based on profile and behaviour.

  • Engagement layer: A Wati or Interakt WhatsApp agent handles all FAQ traffic, reminders and last-minute changes — multilingual, 24/7.


The reported results: a 32% conversion lift and 47% CTR lift from AI-driven segmentation and AI-generated creative, and roughly 13 hours of average weekly time saved per marketer.


The lesson: single-vendor "AI suite" answers are seductive. The real lift comes from one agent per layer, each one good at its job, talking to a shared attendee record. The integration question matters more than the agent question.


Sources: Salesforce Agentforce · RainFocus · Blackthorn


The takeaway across all three


Capability is cheap. Integration is expensive. Guardrails are non-negotiable.

Curious what a similar case study could look like for your events? Get in touch to talk it through.


Sources & further reading

bottom of page