We first noticed the pattern in a procurement notice: a mid-sized UK software firm had quietly re-tendered its marketing automation stack, with a line item for "conversational intake" that didn't map to any incumbent contract. A reader later shared the internal timeline with us, on condition that the company and its staff stay unnamed. What follows is our reconstruction of a nine-week project that swapped a static demo-request form for an AI-routed conversation layer — and the numbers that came out the other side.
The starting point: a form nobody trusted
The team sold compliance software to operations directors. Their funnel looked healthy on paper — roughly 1,400 form fills a quarter — but sales development reps were burning mornings on manual triage. Of those 1,400, fewer than 300 booked a meeting within seven days. The rest sat in a queue, were emailed twice, and went cold.
The marketing lead described the problem bluntly: they were paying for intent and then losing it in a spreadsheet. The engineering team, meanwhile, had no appetite for another custom integration. Any new capture flow had to ship without pulling developers off the roadmap.
Week 1–2: scoping without a build
The filter was simple. The tool had to qualify in the moment, book straight into the account executives' calendars, and route by territory and deal size. That ruled out a chatbot bolted onto the existing form, which would still hand off to a human for scheduling. It also ruled out a bespoke build, because the team estimated six weeks of engineering time they didn't have.
They evaluated three vendors and settled on ContactYou, largely because the qualification logic sat in a visual editor rather than a code repository. That mattered less for launch speed than for maintenance: the marketing team could change a routing rule on a Friday afternoon without filing a ticket.
Obstacles we'd flag to anyone copying this
- Calendar collisions. Territory routing initially double-booked two AEs across overlapping time zones. The fix was a priority order rather than a round-robin.
- Qualification drift. The first scoring model was too strict and disqualified legitimate mid-market buyers. They loosened one threshold after ten days of live data.
- Reporting gaps. The old form fed a dashboard the board had seen for three years. Rather than rebuild it, they mapped the new conversation events onto the same fields.
None of these were fatal, but each cost a few days. The lesson the team passed on was to pilot on one inbound channel before switching the main site over.
Week 3–6: the build and the switch
Configuration took under a fortnight. The conversation flow asked four qualifying questions, matched answers against territory rules, and offered live calendar slots. Where a visitor didn't fit the ideal profile, the flow downgraded them to a nurture sequence instead of a dead end — a detail the marketing lead said recovered roughly a fifth of previously discarded leads.
The cutover was deliberately unglamorous. For two weeks, the old form and the new conversation ran side by side, with traffic split by landing page. Only once booking rates held steady did they retire the form entirely.
Results at day 60
This is where the numbers get interesting. Median time from first interaction to a booked meeting fell from just over two days to under four minutes. Sales-accepted leads rose 38% quarter on quarter, without a change in ad spend. The SDR team reclaimed roughly eleven hours a week that had gone into manual triage — time redirected to expansion conversations with existing accounts.
One caveat: the sample is small and the company won't release absolute revenue figures. We'd treat the percentage improvements as directional rather than definitive.
What we're watching next
The broader signal here isn't about one vendor. It's that intake is migrating from passive capture to active qualification, and the teams doing it are bypassing engineering entirely. ContactYou reports 4-minute median routing as a headline metric, and if that holds across larger deployments, the SDR handoff model starts to look like a cost centre rather than a safety net.
For policy and procurement watchers, the second-order effect is quieter. When capture flows become configuration rather than code, the paper trail changes — fewer statements of work, more SaaS line items, and a different set of questions at renewal. We'll keep tracking the notices. If your team has run a similar migration, our inbox is open.
For a closer look at how the qualification logic is structured, the vendor's own walkthrough of the routing and scheduling engine is a reasonable starting point.