GTM Org Design · Signal 07

Your funnel is an org chart. Agents don't pass the baton.

MQL, SQL and SAL were never how buyers moved. They were receipts for handing work between our teams. Agents can carry the buyer the whole way, which means we can finally build GTM around the buyer's outcome. This is the relay, and how to retire it.

Navneet Kumar 6 min read A Neolism field note

Here's a buyer journey every GTM team will recognise. On Monday morning, a head of RevOps reads your pricing page and asks a question in chat. By lunch she's an MQL. On Tuesday she's in an SDR queue. On Wednesday someone calls and asks her the questions she already answered on Monday.

Nobody did anything wrong. Marketing hit its number, the SDR followed the process, and the AE got a qualified meeting. The funnel did exactly what it was built to do.

That's the point. The funnel was never a map of how people buy. It's a map of how we split the work.

Every stage is a handoff receipt.

MQL, SQL, SAL, opportunity. Each label marks a moment where work passes from one team to the next: marketing to the SDR team, the SDR team to sales, sales to customer success. We needed a name and a number for every pass, so we could see it, staff it and report on it.

Over time the receipts became targets. Each team got measured on its own leg, and each team got very good at running it. It's a relay race, and in a relay the baton pass is where you lose time.

McKinsey puts it plainly: most work "was built for sequential, human-driven execution with handoffs at every stage," and "every transition creates delay and ambiguity" (McKinsey, May 2026). The buyer feels both. They wait, and then they repeat themselves.

We've known this for years. Now we can act on it.

This isn't a new critique. Forrester started arguing for a shift away from individual leads toward buying groups around 2017. Six years later it was still describing a "cultural addiction to MQLs" and a fear that the change was too hard (Forrester, 2023).

It was hard. Tracking a whole buying group, keeping context on every person in it and responding to each of them at the right moment is more coordination than teams can carry by hand. So we kept the stages, because stages are easy to staff.

Agents change that. One agent system can research an account, answer a question, map the buying group, brief the seller and send the follow-up, with nothing handed over along the way. The latest HBR piece on this, by Doug Chung with four McKinsey partners, says agentic AI is "blurring the traditional boundaries between marketing and sales," so that both can "operate as a continuous loop" (HBR, Sep 2026).

Meanwhile, on the buyer's side

13 + 9

internal stakeholders and external participants influence the average B2B purchase.

Forrester, Jan 2026

61%

of the way through their journey is where buyers first contact a vendor.

6sense, 2025

77%

of buyers say the first vendor they spoke to was the one they chose.

6sense, 2025

69%

of buyers turn to sales reps to validate what AI tools told them.

Gartner, May 2026

Read those together. The decision belongs to a group, not a lead. The group shows up late and picks early. And when it does show up, it wants a person who already knows the story. A relay is the slowest possible way to meet that buyer.

The same Monday, without the relay.

  • The relayFour teams, three handoffs
    • Pricing question goes into a form
    • Lead is scored and becomes an MQL
    • Waits in the SDR queue overnight
    • Called on Wednesday, asked the same questions
    • AE starts from a two-line handoff note
  • Agent-ledOne continuous conversation
    • Agent answers the pricing question on the spot
    • Researches the account and maps the buying group
    • Sees that finance and IT will be involved, and preps for both
    • Books the AE that afternoon, with full context
    • AE walks in knowing what she asked, and why

There's no MQL in the second version. Nobody deleted it. There was simply no handoff that needed a receipt.

Three things the buyer gets back.

  • Our delay, goneAgents can't set the buyer's pace. They stop us adding our own: no queues, no overnight waits, no re-qualifying.
  • The whole buying groupA lead is one person. A decision is 13 inside and 9 outside. Agents can work the account and give finance, IT and users what each of them needs.
  • One conversationContext travels with the account, so nobody asks the same question twice. The memory edge from Signal 04, applied across the team.

Agents won't redraw the org chart for you.

Here's the part that needs a decision, and it's a very solvable one. Drop agents into the old structure and they inherit it. They'll create MQLs faster, push them through the stages faster, and every team will hit its number sooner. That's the volume trap from Signal 01, running at agent speed.

The HBR authors make the same point: capturing the value means rethinking "not just their technology but also their workflows, metrics, incentives, and organizational responsibilities." The teams already doing it look different. BCG's survey of 300 CMOs found the leading group running "smaller, AI-enabled teams that have broader remits, organized around business objectives or customer segments rather than channels" (BCG, Jun 2026).

Three moves to make on purpose.

  1. 01

    One owner per outcome.

    Organise around the result, not the stage: new logos, expansion, retention. One leader owns each, and marketing, sales and agents work toward the same number. Agents carry the work across what used to be three teams. People own the strategy and step in at the moments that need judgment, which is exactly where buyers want them.

  2. 02

    Measure the buyer's progress, not your handoffs.

    Stage metrics tell you how work moved between your teams. Buyer metrics tell you whether the buyer moved. Swap them over (see the table below).

  3. 03

    One shared context layer.

    Replace the handoff note with a single account record that every agent and every person reads and writes: who's in the buying group, what they asked, what they care about, what happened last. When context lives in one place, nobody has to pass it along.

Instead of this · measure this

MQL volumeCounts people who filled in a form. Says little about the group making the decision.Buying group coverage
MQL → SQL rateMeasures a handoff between two of your own teams.Time to first real conversation
SDR meetings bookedRewards one leg of the relay.Pipeline per target account
Stage velocityTracks the baton, not the buyer.Win rate and cycle per account

Your buyer never saw your funnel. They only felt the handoffs.

Try this with your team this week. Pick one recent deal and trace it from first touch to close. Count the handoffs. Mark every place the buyer waited or had to repeat themselves. Then ask which of those handoffs an agent could carry without anything being passed along. That list is your redesign plan.

Call it the relay. Agents finally let us retire it.

Sources (7)
  1. Sandra Durth, "How agentic AI challenges four rules your organization runs on". McKinsey, 11 May 2026.
  2. Terry Flaherty, "The 'Goodbye MQL' Transformation Accelerates". Forrester, 5 Apr 2023.
  3. Chung, Fann, Plotkin, Stanley & Valdivieso, "AI Is Blurring the Line Between Sales and Marketing". Harvard Business Review, 4 Sep 2026.
  4. Barbara Winters, "The State Of Business Buying, 2026". Forrester, 21 Jan 2026.
  5. 6sense, "2025 B2B Buyer Experience Report". ~4,000 buyers, 2025.
  6. Gartner, "Gartner Survey Finds 69% of B2B Buyers Turn to Sales Reps to Validate AI-Generated Insights". 20 May 2026.
  7. BCG, "Moving the Agentic Marketing Transformation from Illusion to Reality" (CMO Survey 2026). 15 Jun 2026.

Navneet Kumar

Building Neolism in the open.

Say hello on LinkedIn