Pipeline Leakage: The 8 Places Your Deals Disappear (and the AI Solution)
Pipeline leakage is when deals end up neither won nor lost, they just vanish. The 8 leakage points, what they cost, and how AI counters each one.
Table of contents
What is pipeline leakage?
Pipeline leakage is when deals that enter a pipeline end up neither won nor lost; they simply disappear without being closed. A large share of B2B deals end this way, as a silent bleed in the organisation's sales funnel.
You have a full pipeline. The sales team is busy. The forecast shows a good month. And yet you close less than 50% of what was planned. Deals that looked promising six weeks ago have gone quiet. That isn't bad luck: it's pipeline leakage. A large share of the B2B deals that enter a pipeline end up neither won nor lost, they just... disappear.
Last updated: March 2026
TL;DR:
- A large share of B2B pipeline deals disappear without ending up won or lost
- The 8 most critical leakage points can all be identified and prevented proactively
- AI can detect "deal stagnation" automatically
- The average B2B win rate is 17-22%; the best teams hold 30%+ by closing the pipeline leak
- Each leakage point has a specific AI countermeasure
What is pipeline leakage, and what does it cost your company?
Pipeline leakage is the systematic disappearance of deals from a sales pipeline without a defined ending. Unlike lost deals (where the customer actively chooses a competitor) or won deals (a closed agreement), leakage is characterised by passivity: no activity, no communication, no decision.
Pipeline leakage is measurable financial damage. Let's put numbers on it:
A company with 100 active deals at an average deal value of DKK 150,000 has a pipeline of DKK 15,000,000. With an average B2B win rate of 20%, they should close 20 deals = DKK 3,000,000.
But with typical pipeline leakage of 60%, only 40 deals are actually active. A win rate of 20% of 40 = 8 deals = DKK 1,200,000.
Leakage costs them DKK 1,800,000: 60% of the potential.
According to Gartner's B2B Sales Research, pipeline leakage is the main reason sales teams consistently deliver under forecast: not poor sales technique, not the wrong product, but structural gaps in pipeline management.
The 8 places deals disappear
Pipeline leakage points: cause, frequency and AI solution
| Leakage point | Frequency | Average deal loss | Can AI prevent it? |
|---|---|---|---|
| No next step defined | Very frequent | 35-45% of affected deals | Yes |
| Champion leaves the company | Frequent | 60-70% of affected deals | Partly |
| POC/pilot ran too long | Frequent | 40-55% of affected deals | Yes |
| Multi-stakeholder stagnation | Very frequent | 50-65% of affected deals | Yes |
| Budget freeze | Seasonal | 70-80% of affected deals | Partly |
| Competitor takeover | Medium | 75-90% of affected deals | Partly |
| Ghosting after proposal | Very frequent | 55-70% of affected deals | Yes |
| Decision postponed to next quarter | Frequent | 30-50% of affected deals | Yes |
See also our guide to conversation intelligence for pipeline analysis.
Leakage point 1: no next step defined
It is the most widespread cause of pipeline leakage, and the most preventable. A meeting without a clear, calendar-booked next appointment results in a 3x higher chance of the deal dying than a meeting that ends with a concrete booking.
The reason is simple: when the salesperson and the prospect leave a meeting without an agreement, the initiative is left to the prospect. Most prospects forget, not because they're uninterested, but because they have 50 other priorities.
According to HubSpot's State of Sales Report, only 36% of salespeople consistently define a next step in every meeting. Deals that end with a concrete, calendar-booked next appointment close markedly more often than deals without one. It is the lowest-hanging fruit in pipeline management.
What AI does: AI-based conversation analysis can analyse call recordings and flag meetings that ended without a confirmed next appointment, so that the salesperson gets an alert: "Meeting with [Company] ends without a next step, book now?"
It is automated administration combined with conversation analysis: AI hears the meeting, understands the context and takes action.
Leakage point 2: the champion leaves the company
Every year, 20-25% of all B2B decision-makers change jobs. In a typical B2B sales cycle of 3-6 months, the likelihood that your champion changes role or company is significant.
When your internal champion disappears, three things typically happen:
- The new person doesn't know the background and "resets" the evaluation
- The project's momentum is lost in organisational restructuring
- The new person brings in their own preferred vendors
See also our guide to forecasting and pipeline management.
According to Forrester's B2B Buyer Research forrester.com, salespeople lose 58% of deals in which the primary contact changes jobs during an active sales cycle.
What AI does: The AI system monitors LinkedIn and the CRM for changes in contacts' roles. The system alerts the salesperson within 24 hours and suggests two actions: (1) Identify new stakeholders with LinkedIn Sales Navigator, (2) Send a handover email to the departing champion with thanks and a request for an introduction to their successor.
Leakage point 3: the POC/pilot ran too long
A Proof of Concept (POC) or pilot is a necessary part of many B2B sales processes. But it is also one of the most dangerous leakage points in the pipeline.
The problem is scope creep and open-ended pilots. A pilot without a defined end date and success criteria isn't a pilot, it's free use of your product.
Long, open pilots close markedly less often than short pilots with a fixed end date and clear success criteria. The reason is that the longer the pilot runs, the lower the urgency, and the more time competitors have to come in with counter-offers.
What AI does: Automatic tracking of the pilot start date and an alert at the 30-day mark if no close date has been set. AI automatically generates a draft "pilot review meeting" email with concrete success-criteria questions.
Leakage point 4: multi-stakeholder stagnation
B2B decisions typically involve 6-10 decision-makers. When you communicate with only one or two of them, you risk the decision process stagnating because other stakeholders are never engaged.
Multi-stakeholder stagnation arises when there is consensus resistance internally at the prospect, but the salesperson isn't aware of it. Typical signals: meetings are repeatedly postponed, replies to emails take ever longer, "we're nearly ready" is repeated over months.
Deals with several actively engaged stakeholders close markedly more often than deals with only one or two contacts.
What AI does: The AI system tracks engagement breadth: the number of unique contacts who are active in the deal dialogue. A deal with 8 stakeholders at the company but only 1 engaged contact is flagged automatically as "Stakeholder Gap Risk."
Leakage point 5: budget freeze
Budget freezes hit unexpectedly and are rarely predictable at deal level. But across the pipeline it is possible to identify patterns.
Budget freezes are typically seasonal (Q4, start of Q1) and clustered by industry. If three deals in the same industry simultaneously go quiet in November, the likelihood is high that it's a budget problem, not a product problem.
AI can identify "budget freeze clustering": situations in which many deals in the same industry or geography stop moving at the same time. It is a signal of market-level challenges, not deal-specific ones.
What AI does: The system identifies clusters of stagnating deals and classifies them. For budget freeze clusters, AI recommends: (1) Put deals on "Nurture" status rather than Active, (2) Shift effort to new prospects, (3) Plan proactive outreach for the Q1 budget opening.
Leakage point 6: competitor takeover
Competitors rarely steal deals with better products. They steal them with better timing and better relationships. The classic scenario: you are in late dialogue with a prospect, and suddenly a competitor is introduced by a contact with an internal network you didn't know about.
Competitor takeover is almost always a multi-stakeholder problem: you communicated with the wrong person, and the competitor found the right one.
What AI does: Conversation intelligence identifies when competitors are mentioned in calls or emails. The system alerts the salesperson with: "Competitor X mentioned in customer dialogue, recommended action: senior executive engagement and ROI documentation."
According to Forrester Research, executive-to-executive engagement increases the win rate by 23% in deals where competitors are actively involved.
Leakage point 7: ghosting after proposal
It's every salesperson's nightmare: you send a carefully prepared proposal, and the prospect vanishes. No reply to emails. No response by phone. Silence.
Ghosting typically happens for two reasons:
- Internal chaos: The prospect has an internal change of priorities and doesn't have the time or the desire for a difficult conversation
- Hidden rejection: The prospect has decided on no, but doesn't want to confront the salesperson
A large share of all proposals never receive a definitive answer, either yes or no. It is pipeline leakage in its purest form.
What AI does: Automatic tracking of the proposal date. If there is no response within 5 working days, AI generates a follow-up template specific to the proposal. If there is no response within 15 days, the system flags the deal as "stalled" and recommends a change of tone in outreach: from sales-focused to helpful ("What could we do better in our proposal?").
The proposal date is logged automatically in the CRM when it is sent, so tracking requires no manual action from the salesperson.
Leakage point 8: decision postponed to next quarter
"We're interested, but it fits better in next quarter's budget." It is the most sophisticated form of ghosting, because technically the prospect isn't saying no.
Quarter postponement is a deal killer in 30-50% of cases. The reality is that if the priority isn't high enough to act on now, it is rarely higher next quarter, and you are another quarter away from close.
What AI does: The AI system flags all deals with "Next Quarter" as the close date and calculates an "urgency score" based on the historical conversion rate for quarter-postponed deals. The salesperson receives an alert with the recommendation: create urgency now, or accept that the deal is probably lost.
The best salespeople are the ones who understand the difference between a genuine quarter postponement (budget cycle) and a polite no (indifference). AI helps to tell them apart based on engagement data.
How AI stops pipeline leakage
Pipeline leakage is primarily a visibility problem. Salespeople and sales managers can't act on data they don't have. An AI-driven approach solves this with four layers:
Layer 1: Signal Collection: All deal activities are tracked automatically: email opens, calendar invitations, calls, proposal opens, LinkedIn activity from key stakeholders.
Layer 2: AI Analysis: The algorithm calculates a "Deal Health Score" from 0-100 for each deal based on activity level, engagement breadth, time since last contact and historical win-pattern match.
Layer 3, Prediction Engine: The machine learning model predicts close probability based on hundreds of historical deals. A deal with a Health Score below 40 for more than 14 days is flagged as "At Risk."
Layer 4, Alert System: Salespeople and sales managers receive daily "Pipeline Health" notifications with concrete action instructions for each at-risk deal.
The result: the sales team no longer spends the bulk of its time on CRM updating and manual pipeline review; it spends it on the specific interventions AI has identified as most critical.
Pipeline hygiene as a cultural foundation
AI and technology solve the visibility problem. But pipeline leakage also has a cultural dimension: a sales culture that tolerates "zombie deals" in the pipeline.
A zombie deal is a deal that is technically alive but clinically dead. No activity for 30+ days, no defined next step, but the salesperson doesn't take it out of the pipeline because it affects the quota attainment figure.
The consequence: sales managers forecast on the basis of a pipeline that is 30-40% inflated with zombie deals. Forecast accuracy falls to 40-60%, and decisions are made on the basis of faulty data.
Pipeline hygiene requires three cultural agreements:
- Honesty: Salespeople mark deals as lost when they are genuinely lost, not "on pause"
- Discipline: No deal lives in the pipeline without a defined next step
- Cadence: Weekly pipeline review meetings with a data-based conversation, not gut feelings
AI supports pipeline hygiene by automatically marking deals that have had no activity for 21+ days as "Stalled" and requiring the salesperson's confirmation: "Is this deal still active?"
Integration with win-loss analysis
Pipeline leakage is inseparable from win-loss analysis. Understanding where deals leak requires systematic analysis of why they leak, and that requires data from both won and lost deals.
Pipeline data should be linked with win-loss patterns: which leakage points are most widespread in your specific industry, with your specific competitors, at your specific price level?
Over time, AI builds a "leakage fingerprint" for your pipeline: the specific combinations of signals that predict deal death in your context. It is far more precise than generic B2B benchmarks.
For a deeper analysis of what happens when deals are lost (and what they tell you about your sales strategy), see our guide: Win-Loss Analysis: What Lost and Won Deals Tell You About Your Strategy.
Pipeline velocity: the connected metric
Pipeline leakage has to be understood in relation to pipeline velocity, the speed at which deals move through your pipeline.
Pipeline Velocity = (Number of deals × Average deal value × Win Rate) / Average sales cycle (days)
Leakage effectively reduces all four parameters: fewer active deals, lower average value (zombie deals drag it down), a lower win rate and, paradoxically, lengthens the sales cycle, because resources are spent on dead deals rather than active ones.
Read more in our guide: Pipeline Velocity: How to Accelerate Your Sales Speed with AI.
AI and GDPR in pipeline tracking
Pipeline tracking via AI involves email analysis, call monitoring and LinkedIn monitoring. All of these activities have GDPR implications.
For AI-driven pipeline tracking in Denmark, the following applies:
- Email tracking (open pixels, click tracking): Doesn't require separate consent, but must be mentioned in the privacy policy and in the email footer
- Call recording: Requires explicit consent from the prospect before recording starts (the Danish Marketing Practices Act + GDPR)
- LinkedIn monitoring: Use only publicly available data (job changes, etc.), not scraping
- Data storage: All pipeline data is processed and stored within the EU
For a complete guide to the legal use of AI in sales, see AI Calls and Legislation in Denmark 2026.
Proactive pipeline management: a 4-week plan
Week 1: Audit: Go through all active deals and classify them: Active (activity within 14 days), Stalled (14-30 days), Zombie (30+ days). Remove or mark all zombies as lost.
Week 2, Prioritisation: Rank the remaining deals by Deal Health Score and identify the 20% with the lowest score. These are your immediate intervention priority.
Week 3, Playbooks: Define specific playbooks for each leakage point: what is the right action for a deal with ghosting? For a champion who has left? For an overrun POC?
Week 4, Automation: Implement AI-based tracking and alerting. From week 5 the sales team spends a maximum of 30 minutes a week on pipeline review; the rest is AI-monitored.
Want help implementing automatic pipeline management? See also JesperAI, which works with AI calls.
Frequently asked questions about pipeline leakage
Q: What is a normal win rate in B2B sales? The average B2B win rate is 17-22% of all deals that enter the pipeline (including leakage). For deals that reached the proposal stage, the win rate is typically 25-35%. The best sales teams in SaaS and professional services hold a 30-40% overall win rate through systematic pipeline hygiene and AI-assisted deal management.
Q: When should you close a deal as lost rather than stalled? A deal should be closed as lost if 30+ days have passed without a response despite 2+ outreach attempts, or if the prospect has explicitly said "not now" without a concrete future date. Keeping deals artificially open ruins forecast accuracy and wastes resources. An honest pipeline is always better than an inflated one.
Q: Can AI predict whether a deal will succeed from the start? AI can calculate win probability based on early-signal matching against historical winning deals. Signals such as company size, the decision-maker's title, the number of engaged stakeholders and industry are statistically correlated with win rate. But prediction is always probabilistic, not deterministic. AI improves forecast accuracy from ~50% to ~75-80%, but perfect prediction doesn't exist.
Q: What is the difference between pipeline leakage and a low win rate? A low win rate refers to the percentage of deals you lose in direct competition. Pipeline leakage refers to deals that disappear without a clear cause (ghosting, stagnation, no decision). A company can have a relatively high win rate (30%) but enormous pipeline leakage, and thus still underdeliver on revenue. Both problems require separate solutions.
Q: Which CRM systems support AI-based pipeline leakage tracking? HubSpot, Salesforce and Pipedrive all support AI-based deal health scoring via native features or integrations.
Conclusion: pipeline leakage is a solvable problem
Pipeline leakage isn't a law of nature, it's a structural weakness in pipeline management that can be eliminated with the right processes and the right technology layer.
The 8 leakage points are all identifiable, measurable and actionable. AI makes it possible to monitor all 8 simultaneously, for all deals, without requiring extra resources from the sales team.
The result: a pipeline that actually reflects your real business situation, and a sales team that spends its time on the deals that can be won, not on tending zombies.
- See the connections in our Pipeline Management AI Guide
- Understand Pipeline Velocity and sales speed with AI
- Read our Win-Loss Analysis guide
- See our Win-Loss Interview Guide for practical questions
Book 20 minutes of AI consultation with Agent360, and talk about how AI can help you hold on to your deals.
For an automated top of funnel: see also JesperAI on automatic meeting booking.
#pipeline leakage #pipeline management #AI #B2B sales #deal management
Contents
Should your sales reps spend their time talking?
We have built it for our own sales: the call is transcribed, the AI suggests the outcome, and the email is ready. We look at how your sales run today.
Book a reviewShould your sales reps spend their time talking?
We have built it for our own sales: the call is transcribed, the AI suggests the outcome, and the email is ready. We look at how your sales run today.
How we built it · Sales & follow-up
Sales cockpit with AI
A call list for salespeople. The call is transcribed in Danish, AI suggests the outcome (e.g. “Call again”) and drafts the email based on what the customer said.
- The salesperson calls
- AI writes transcript and email
- The salesperson sends
Related articles

MFA: multi-factor authentication for your business – how to get started
Learn how to implement MFA securely in your organisation. Guidance on methods, costs and best practices for protecting company data.

English-language invoices: requirements for invoices to foreign customers
Invoice in English to foreign customers legally. Danish regulations, EU and international VAT rules, zero-rating, reverse charge, and automation.

Public Procurement in Denmark: how to bid for government contracts
How Denmark's Public Procurement Act works for suppliers. Threshold values, tender procedures, exclusion grounds and AI solutions for proposal management.