Questex, a New York-based B2B media and events company, deployed two AI sales agents built by 11x and closed $1.056 million in attributed revenue within 90 days. The inbound agent — named "Julian" — calls every new lead within two minutes of form submission. Outbound agent "Alice" generated an additional $1 million in new pipeline over the same period. Every dollar was tracked through Salesforce and Gong.
The speed-to-lead problem Questex solved
The business case here is deceptively simple: respond faster, close more. Research from MIT's Lead Response Management Study (replicated multiple times through 2024) shows leads contacted within one minute convert at 391% above baseline. Leads contacted within five minutes are 21 times more likely to qualify than those contacted after 30 minutes.
And yet, the average B2B company still takes 47 hours to respond to an inbound lead. Sixty-three per cent never respond at all. Questex was slightly better than average but still losing deals to the gap.
"We hit a breaking point on sales bandwidth due to our rapid revenue growth," said Rhiannon James, Group President at Questex. "Our marketing team was generating more inbound leads than our SDR team could handle."
How the two agents actually work
Questex runs a dual-agent architecture. Each agent owns a distinct phase of the sales funnel:
Julian (inbound) monitors form submissions. When a prospective exhibitor expresses interest in a trade show, Julian calls within two minutes, asks qualifying questions, and routes the lead to the appropriate human account executive. Leads that don't immediately convert enter automated follow-up sequences through Gong Engage Flows.
Alice (outbound) operates as a fully autonomous prospecting engine. Questex defines ideal customer profiles by event, industry, and role. Alice sources contacts, enriches them with public data, builds personalised multi-touch outreach sequences, handles replies, and books meetings directly onto sales calendars — syncing everything to Salesforce.
"We started where the pain was sharpest and the process was most repeatable: outbound pipeline generation for our events business," James explained. "We were iterating slowly, not because of strategy, but because of setup time."
The attribution model that makes this credible
What separates Questex's results from the typical "AI saved us millions" hand-waving is the measurement discipline. The team tracks performance across every funnel stage — percentage of leads contacted within the target response window, conversion to meetings, meetings to opportunities, opportunities to close.
Revenue attribution runs through Salesforce and Gong. Any opportunity where the first meaningful touch came from an AI agent gets tagged accordingly. "We measure this rigorously," James said.
The results after three months:
- $1.056 million in closed revenue attributed to Julian (inbound)
- $1 million+ in new pipeline generated by Alice (outbound)
- Meeting conversion up from 30% to 37%
- Response time down from hours/days to under two minutes
Where they drew the line
Questex was deliberate about what AI should not touch. Human account executives handle all discovery calls and closing conversations. AI never makes first contact with strategic or high-value accounts, existing client relationships, late-stage deals, or categories where tone matters more than speed.
"AI is not a universal first touch," James said.
The early deployment also surfaced problems. Some prospects could tell the outreach felt automated and disengaged. Others asked outright whether they were talking to a bot. Questex responded by tightening personalisation prompts and sharpening qualification criteria using Gong call data to define what a good-fit prospect actually looks like.
The workforce question — answered practically
The SDR team initially feared replacement. Questex responded by giving existing reps ownership over the AI program itself. Today, several SDRs manage quality control, monitor campaign performance, and adjust outreach sequences. The repetitive setup work — list building, enrichment, sequence construction — is handled by Alice and Julian. The humans focus on judgment calls.
"Once reps saw the results, adoption accelerated," James said.
This mirrors broader market data on hybrid configurations. Research from 2026 shows that hybrid pods — one human SDR per two AI agent seats — out-book pure AI configurations by 1.9x per dollar spent, and outperform human-only pods by 2.4x. The optimal model isn't replacement; it's augmentation with clear boundaries.
Why this matters right now
The AI SDR market hit $5.81 billion in 2026, growing at 32.3% CAGR according to Research and Markets. Forty-one per cent of enterprise B2B teams now run at least one AI SDR in production — up from 12% one year earlier. This is no longer early-adopter territory. It's becoming table stakes.
For Australian business owners running 10-50 person teams, the maths is worth examining. A fully loaded human SDR in the US costs $98,000-$173,000 annually. AI agent programs run $6,000-$30,000 per year. Even accounting for the "show rate penalty" — AI-booked meetings have slightly lower attendance than human-booked ones — the unit economics are compelling if your inbound volume exceeds what your current team can handle.
The prerequisite is the same one Questex's Rhiannon James emphasised: "Fix your process before layering AI on top. AI amplifies whatever you plug it into. If your ICP definitions, CRM hygiene or follow-up processes are messy, AI will scale the mess."
This echoes what we've observed in the rise of AI agent marketplaces — the companies seeing real results aren't the ones rushing to deploy every agent available. They're the ones with clean processes who use AI to remove specific, measurable bottlenecks.
What to watch
Questex is now expanding AI-assisted production across its events portfolio, using AI for video scripting and generation at one-tenth of previous costs. The company produced a promotional video for its Bar & Restaurant Expo in 24 hours for approximately $2,000 — a project that previously took weeks and cost ten times more.
The broader signal: Gartner predicts 40% of enterprise applications will feature task-specific AI agents by end of 2026, up from less than 5% in 2025. If your sales process involves any repeatable, high-volume top-of-funnel activity — and your response time is measured in hours rather than minutes — the Questex case study suggests the ROI timeline is weeks, not years.
Sources
- Questex Used AI Agents to Close $1 Million in 90 Days — A Media Operator
- AI SDR Market Report 2026 — Research and Markets
- Speed to Lead Statistics 2026 — GreetNow
- Speed to Lead in 2026: Why Response Time Still Wins — SpurIQ
- AI SDR Statistics 2026: 100+ Outbound Sales Data Points — Digital Applied
- This AI Shift Is Starting to Replace a Key Business Function — TheStreet
