Part 3 of the Formless Funnel series. Part 1 defined the model. Part 2 showed why teams are switching. Now we’ll get into the mechanics: how the conversation starts, how qualification happens, how routing works, and what a real handoff to a rep looks like.
Most teams go formless and immediately hit the same wall: the conversations start, but the pipeline doesn’t follow. Usually it’s not the idea that’s broken. It’s the mechanics.
None of this was possible before AI. The old funnel was built around a constraint: you couldn’t have a real conversation with every visitor, so you took their info and followed up later. AI removed that constraint. The formless funnel is the architecture that follows.
Capture first, then qualify
Forms were a workaround. You couldn’t talk to every visitor in real time, so you collected what you could and followed up later. Email became the default follow-up channel because it scaled. The cost was conversion: follow-up email reply rates sit around 9%, while direct messaging clears 83%. Most of the pipeline you’re missing lives in that gap.
AI removed the constraint. A formless funnel captures through a direct messaging entry, not a form, and the thread stays open from there. Qualification happens inside the thread instead of in front of it. Everything downstream, routing, enrichment, handoff, runs in the same place the buyer first showed up.
This is where customer acquisition cost actually changes. CAC isn’t only about ad spend. The compound cost is every dropped conversation, every cold follow-up that didn’t get a reply, every SDR hour spent chasing a buyer who already moved on. Messaging closes that gap because the thread doesn’t end. When intent dips, the thread stays open and re-engagement costs almost nothing.
Qualify inside the conversation
The most common mistake is replacing the form with qualification questions dressed up as chat. That’s still a form. The interface changed, but the logic didn’t.
A formless conversation starts from a different assumption. The buyer landed on a specific page for a reason, and the opening’s job isn’t to extract data. It’s to meet them where they already are. On a pricing page, that means acknowledging what they’re trying to figure out. On a G2 reviews page, it means helping them evaluate. On a demo CTA, it means getting them to the demo, not making them justify why they deserve one first.
The qualifying questions still get asked. They get woven in after there’s been an exchange of value. The difference between “Can you tell us your team size?” as an opening line and “Are you thinking about this for a larger team or starting smaller?” mid-conversation is enormous. One reads like a screening. The other feels like someone paying attention.
A few principles hold the design together:
Lead with what the buyer’s already thinking about. A buyer on pricing is thinking about cost and what they get for it. A buyer on integration docs is thinking about fit with their existing stack. The opening should feel like the conversation was already in progress.
Let identity detection and enrichment do the work. Real-time enrichment pulls role, company size, industry, and tech stack in the background from the first interaction. By the second exchange, you already know who you’re talking to. That changes what you ask next.
Ask one thing at a time. The fastest way to make a conversation feel like a form is to fire multiple questions at once. Ask, get a response, learn from it, ask the next thing that follows.
That’s where the no-show improvement comes from. For example, one Knock AI customer’s no-show rate dropped from 32% to under 3% after moving to chat-based discovery. When qualification happens inside the thread, buyers show up because they’ve already gotten value and have a reason to be on the call.
Route as you learn
Qualification used to be a one-time event. The form fired, the lead scored, the routing rules kicked in, and the buyer landed somewhere. Whatever happened after that didn’t change the routing decision.
In a formless funnel, qualification is continuous. The system keeps learning as the conversation moves, and routing adjusts in real time. A buyer who looked like a low-priority browser in message one might reveal themselves as a high-fit decision maker by message four. The routing shouldn’t be locked into the first read.
Several decisions are running in parallel, and getting revisited as new information comes in.
Identity and firmographics. Who’s talking, and what company are they from? Real-time enrichment runs from the first interaction.
Fit against ICP rules. Fit scoring runs against company size, industry, revenue, geography, and tech stack. A strong match routes differently than a marginal one.
Territory and CRM ownership. Territory rules apply immediately, so the right person gets the conversation, not just whoever’s available.
Rep availability. If the owning rep is live, the conversation goes to them directly. If not, the system chooses: queue for the rep, route to a backup, keep the AI running, or offer a booking.
Intent level. A buyer who landed on pricing and immediately asked about annual contracts reads differently than someone clicking around a general blog post.
The point isn’t to involve humans more. It’s to involve them only when it actually moves the deal forward. This is what AI changed. The old funnel had to send every conversation through a human gate because there was no other option. Now there is. AI handles entry, qualification, and enrichment. Reps come in when there’s a buyer worth their time and a question worth their answer. Every other path stays inside the conversation.
Every scenario gets a defined path, set in advance. Low-fit visitors get what they need from AI, and rep calendars stay clean. After-hours high-fit visitors get a real booking option with the right rep, not a generic dead end. Reps don’t triage manually. The system decides fast, because a buyer who’s ready to talk right now won’t stick around.
Stay in the thread
The most important rule of the handoff is the simplest: don’t switch channels.
Most inbound funnels break exactly here. A buyer has a solid conversation with an AI or an SDR, then gets pushed to a calendar invite, an email confirmation, and a Zoom call the next morning. Three channels, three context resets, three chances for the buyer to lose interest.
McKinsey’s 2024 B2B Pulse Survey found that 54% of buyers would abandon a purchase or switch suppliers after a poor omnichannel experience, specifically, failing to let buyers move between channels without losing context or having to repeat themselves. The channel switch is the failure mode. The cold handoff is just the most common version of it.
In a well-built formless funnel, the handoff stays in the thread. The rep joins the same conversation the AI started. The buyer doesn’t get re-introduced or re-qualified. Nothing signals they’ve been transferred.
What makes that possible is that the context travels with the thread. By the time a rep joins, they have the buyer’s name, role, and company, which page the buyer came from, what the buyer said with qualification flags surfaced, and CRM context on the account, owner, and any open opportunity.
The rep’s first message moves things forward. Something like: “Looks like you’re evaluating us for your enterprise SDR team. Happy to discuss what that would look like for your setup.” That’s only possible if the record traveled, the rep read it before joining, and the buyer didn’t get bounced to a different channel in between.
The same logic applies to scheduled meetings. When a booking comes through a formless flow, the calendar invite includes enrichment data, the conversation transcript, and intent signals. The discovery call picks up the conversation that’s already happened.
How it works end to end
A VP of Sales at a SaaS company lands on your pricing page. She’s spent time on G2 comparing you to two competitors and clicks “Talk to us.”
The opening message acknowledges she’s on pricing and asks what she’s trying to figure out. She’s evaluating the product for her inbound SDR team. Within two exchanges, real-time enrichment has confirmed her role, company size, and tech stack. Strong ICP match.
The AI asks two follow-on questions that feel like natural next steps. Team size. Current process. By the third exchange, qualification is done, and the routing logic has updated her priority twice based on what she’s said.
It’s 9 PM her time and the owning SDR is offline, so the system routes to a booking flow with the right AE, filtered to her time zone, with the transcript and enrichment summary attached to the invite. She books in the same thread she started in.
The AE joins the next morning already knowing the company, the use case, the team size, and what she was trying to figure out. The call starts in the middle, not at the beginning. Under two minutes from first click to booked meeting, with no forms, no channel switches, and no context lost.
The playbook: start with one entry point
Pick one page, get the conversation right, then build the routing around it.
Step 1: Open the capture channel. Decide where the thread will live before you write a single conversation flow. Direct messaging entry, not a form. That decision shapes everything downstream.
Step 2: Design the conversation for one high-intent page. Your pricing page or primary CTA. Write the opening message for that page specifically. Start with what the buyer’s thinking, not with what you need to know. Then identify four or five qualification questions that actually predict fit and decide how they’ll surface naturally.
Step 3: Build the routing decision tree before going live. Map every scenario: high-fit and rep available, high-fit and rep offline, medium-fit, low-fit, after-hours, non-ICP. Every undefined scenario becomes a place the buyer falls out.
Step 4: Set the handoff standard in writing. Stay in the thread. Decide exactly what context travels: enrichment data, conversation transcript, intent flags, CRM match. If a rep joins without that context, or if the buyer gets bounced to a different channel, the handoff failed.
Step 5: Confirm rep readiness. Good routing still breaks if reps aren’t checking context before they join. Run a short internal session showing what a complete handoff looks like and what reps should do with it before their first message.
Step 6: Measure two numbers for 30 days. Time from first message to live conversation or booked meeting, and the percentage of those that become pipeline within two weeks. Those two numbers tell you whether the mechanics work and where to fix them.
Once the mechanics are running, a different set of questions shows up. What does this mean for how reps spend their time? How do you measure a team that’s no longer chasing form fills? That’s what the next piece will cover: What actually changes inside the org when the formless funnel is working the way it’s supposed to.
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.png)
.svg)
.svg)
.svg)
