I build the revenue system behind B2B SaaS scale.

I recently spoke on a panel at CRO Exchange, San Francisco, June 2026, about one of the most important operating questions in revenue leadership: how do you create a pipeline that predictably converts into revenue in the quarter you need it?

A broken pipeline is not always obvious. Sometimes the CRM looks complete, the deal is in the right stage, the close date is in the quarter, and the rep has a compelling story. But the buyer's behavior tells a different story. That is the signal I look for first.

When the seller's narrative does not match the buyer's behavior, pipeline integrity has slipped. A seller may say the deal is progressing, but has the customer confirmed the decision date? A seller may say the opportunity is a commit, but is procurement engaged? A seller may say the buyer is excited, but has the economic buyer agreed to the business case?

Pipeline that converts into revenue has observable customer evidence. Pipeline that does not convert often depends too heavily on rep interpretation. That does not mean reps are being dishonest — they are enthusiastic and have optimism. But forecast optimism is not the same as forecast accuracy.

A real in-quarter opportunity should have a clearly identified business problem, a measurable impact, a buying event, an economic owner, an agreed decision process, and a reverse timeline that supports the expected close date. The close date should not be a hope date. It should be the output of the buyer's process.

This is especially important because revenue predictability is not only about bookings. A signed contract may still need to move through provisioning, onboarding, service start, usage, acceptance, or other steps before revenue is recognized. That is where many pipelines break: • They track sales activity, but not buyer commitment • They track stage progression, but not decision readiness • They track close dates, but not the operational path to revenue • They track pipeline volume, but not pipeline conversion quality

Predictable revenue is created when Marketing, Sales, Rev Ops, Customer Success, and Finance share one view of what "real pipeline" means. Pipeline integrity is about creating the operating discipline to know earlier. The goal is not to make the CRM look better. The goal is to make the business more predictable.

One of the most important points I made at CRO Exchange: a stage is a claim about buyer commitment. Exit criteria are the evidence required to make that claim.

Too many companies define pipeline stages around seller activity. A meeting happened. A demo was completed. A proposal was sent. Those activities matter, but they do not prove the buyer is moving toward a decision. A demo completed is an activity. The agreed-upon success criteria are evidence. A proposal sent is an activity. A customer-owned next step is evidence. A verbal "this looks great" is enthusiasm. A documented decision process with dates is evidence.

Stage discipline should answer a specific question: what has the buyer done that proves this opportunity belongs here? Not what has the seller done. Not what do we hope happens next. What has the buyer committed to?

False progression is one of the biggest threats to forecast accuracy. It happens when an opportunity moves forward because the seller completed an action, not because the buyer increased commitment. The deal appears to advance, pipeline value increases, forecast confidence improves — but the underlying buyer evidence has not changed. That is how companies end up with pipeline that looks strong in week four and collapses in week eleven.

At one company, we made pipeline accuracy a shared management responsibility. Marketing, Rev Ops, and Sales spent dedicated time together to create and approve pipeline criteria and stages. Marketing understood source quality and intent patterns. Sales understood buying committees and deal progression. Rev Ops understood conversion rates and stage aging. When we brought those views together, the stage criteria became much stronger.

The manager's job was to work with the seller to decide whether to validate the close date, reset it to a more realistic date, or remove the opportunity from the active forecast. We rewarded accuracy and early correction, not optimism. Leaders get the behavior they reward. If they reward optimism, they get optimistic forecasts. If they reward evidence, they get predictable revenue.

AI is being discussed everywhere in revenue organizations right now. Some of that discussion is useful. Some of it is hype. When it comes to pipeline predictability, I believe AI can absolutely move the needle — but it does not replace process discipline, data hygiene, or leadership accountability.

For pipeline integrity, I see the most valuable AI use cases in three areas: capture, corroboration, and prioritization.

Capture means turning revenue activity into structured signals. Emails, meetings, calls, notes, and transcripts contain valuable information that is often scattered across systems. AI can extract signals such as stakeholders, objections, next steps, dates, decision criteria, sentiment, competitor mentions, and procurement timing. But capture alone is not enough — more data can create more noise if the team has not defined which signals matter.

Corroboration is where AI becomes much more powerful. The CRM may say an opportunity is in Commit, but AI can analyze underlying interactions and flag that there is no confirmed decision maker, no recent executive engagement, no procurement activity, no legal review. That is an important disconnect. The point is not for AI to make the forecast call — the point is for AI to expose the gap between what the opportunity record says and what the buyer behavior shows.

Prioritization helps frontline managers focus. AI can identify: which close dates are unsupported, which late-stage deals have gone quiet, which opportunities have no executive engagement, which deals resemble past pushed accounts, and which opportunities resemble past closed-won deals.

In one organization, I worked with our Rev Ops team to compare closed-won deals against pushed accounts from the same period. We used AI to detect patterns across both groups. The result was an 8% improvement in predictability — which helped Finance plan with more confidence, helped Sales leaders coach earlier, and helped Marketing understand which demand was converting into real opportunities.

The biggest mistake is using AI to automate bad assumptions. If stage definitions are weak, AI will scale weak stage definitions. If managers are rewarded for optimism, AI will surface risk that no one wants to act on. AI works best when the organization has the discipline to define what good looks like and the courage to act when the signals are uncomfortable.

Pipeline cleanup sounds tactical — CRM hygiene, close-date updates, stage corrections. But done correctly, pipeline cleanup is much more than that. It is a reset of the organization's definition of truth.

The fastest way to create fake pipeline is to punish people for telling the truth. That is why pipeline cleanup has to separate deal quality from seller quality. A deal that needs to be moved out is not automatically a bad seller. A close date that needs to change is not automatically a failure.

If sellers believe they will be punished for surfacing reality, they will protect the appearance of pipeline. If managers believe their job is to protect the pipeline number, they will allow weak opportunities to stay in the forecast too long. Pipeline cleanup cannot be a blame exercise. It has to be a business accuracy exercise.

After one difficult quarter where revenue goals were missed, we gave the team a one-time amnesty. The message was clear: we are going to reset the pipeline based on buyer evidence. This is not about blaming people for where the pipeline is today. This is about creating a clean baseline so we can make better decisions going forward.

Pipeline cleanup is not about shrinking the pipeline for the sake of being conservative. It is about distinguishing between pipeline that can convert now and pipeline that may convert later. In-quarter pipeline needs urgency, buyer commitment, clear next steps, and a feasible path to revenue recognition. Future-quarter pipeline may need nurturing, education, or timing alignment. When those two pipeline types are blended, the forecast becomes unreliable.

After the cleanup, we baselined the new pipeline and installed a weekly inspection rhythm supported by AI. A one-time cleanup can make the dashboard look better. A management system keeps the pipeline honest. The weekly rhythm focused on: which deals changed stage, which close dates moved, which opportunities lost engagement, which deals had next steps that slipped, and which opportunities resembled previous pushed accounts.

Pipeline cleanup should produce three outcomes: 1. A cleaner view of what can realistically convert into revenue in the current quarter 2. A better understanding of what needs to happen to improve conversion 3. A repeatable management rhythm that prevents the same problems from rebuilding Without the third outcome, cleanup becomes a recurring fire drill. Cleanup is an event. Integrity is a management system.