Why the "What's Wrong With Our SDRs?" conversation keeps happening.

I see this pattern a lot in GTM reviews. The issue is not always demand generation. It is demand progression.

Demand generation is the upstream activity. Campaigns. Events. Inbound. Outbound. Product signals. Partner activity. Intent data. Most teams are not short of signals. If anything, they have too many.

The problem starts after the signal is created.

What should happen next? How quickly should it happen? What counts as pickup? What proves progression? What happens when nobody acts? What gets escalated? What gets quietly ignored because everyone assumes someone else has it?

That is where pipeline leaks. Not always at the top of the funnel. Not always because the campaign was poor. Not always because SDRs were slow. Sometimes the issue is much more boring. And in GTM, boring issues are usually the expensive ones.

Events are brilliant at exposing progression gaps.

Same event. Same booth scans. Same campaign source. Same follow-up expectation. Completely different pipeline outcomes across teams or regions.

The first question is usually: "Did the event work?" Fair question. But often too early. A better question is: "Did we have a consistent system for what happened after someone showed up?"

Because if one region treats an event attendee as priority follow-up, another puts them into nurture, another waits for the AE to decide, and another never assigns them properly, you are not measuring event performance. You are measuring operational inconsistency.

That is not an event problem. That is a progression problem that events happened to reveal this week. Next week it might be inbound. The week after that, outbound. Different channel. Same gap. Same meeting. Same spreadsheet. Possibly the same biscuits.

There are four things I look for when diagnosing this.

1. Same signal. Defined response. Every time.

Not usually. Not when the team has capacity. Not when someone remembers the follow-up rules from a slide created 18 months ago. Every time.

If the same signal produces ten different responses depending on rep, region, segment, or source, you do not have a scalable GTM process. You have a set of personal interpretations — and to be fair to the reps, most are doing what they think is right. The issue is usually not effort. The issue is that nobody has clearly defined what "right" looks like.

2. One team owns each stage

Multiple teams can engage. Marketing can influence. SDRs can prospect. AEs can work. Customer teams can engage. Partners can introduce. But only one team should be accountable for moving that person or account forward at any given stage.

This sounds obvious. It is not how many systems actually work. A lead gets assigned. Coverage looks good. The dashboard says everything is fine. Three weeks later, nothing has moved. Why? Because assignment was treated as ownership. It is not. An owner name on a record is just a name attached to a problem — a very well-formatted problem.

3. Pickup must be proven by movement

This is the one I keep coming back to. Ownership is not pickup. Assignment is not pickup. Coverage is not pickup.

Pickup means something changed. A status moved. A qualification happened. A disposition was captured. A meeting was booked. A nurture decision was made. A clear next step exists.

If a record has an owner but sits in the same status for two weeks, it has not really been picked up. It has been parked — possibly in a very expensive CRM car park.

4. SLAs must be enforced, not documented

Most companies have SLA documentation. It is usually in a slide. The slide was probably created after a workshop. The workshop probably felt productive. Someone may even have said "this is exactly what we needed." Then the deck went into a shared folder and quietly retired.

That is not an SLA. That is a historical artefact. A real SLA is measured. A real SLA has ownership. A real SLA has breach logic. A real SLA gets surfaced before the quarter is already lost. If nobody measures it, nobody manages it.

AI can absolutely help. AI-assisted routing can reduce response times. AI scoring can help prioritise large volumes of signals. AI enrichment can improve context. AI agents can help teams move faster than they could manually.

But AI does not fix a broken progression model. It accelerates it.

If your denominator is wrong, AI learns from the wrong denominator. If ownership is unclear, AI routes into ambiguity. If lifecycle stages mean different things to different teams, AI will automate the confusion with impressive confidence.

That is when the archaeology begins. And nobody enjoys CRM archaeology.

Before asking: "Why are SDRs not following up?" Ask: "Are we measuring them against the real workable pool?"

Before asking: "Why did this campaign not convert?" Ask: "Did every signal receive a consistent next action?"

Before asking: "Why is pipeline below target?" Ask: "Where exactly is progression breaking?"

Is it routing? Assignment? SLA breach? Sales acceptance? Lifecycle definition? Status movement? Opportunity creation? Stage velocity?

Each one has a different cause. Each one needs a different fix. And AI applied to the wrong constraint will not solve the issue. It will just create more volume at the broken point.

AI does not create pipeline. Consistent action on the right signals does.

AI helps you do that faster, at a scale humans cannot manage manually. But the sequence matters.