ICYMI: What AI Won’t Reveal About Your Customers (Recap of Our Pulse Presentation)
At Gainsight’s recent Pulse conference, IcebergIQ Co-Founder and CEO Natasha Narayan and Customer Revelations Founder and Chief Advisor Ruben Rabago took the stage to discuss a topic that is increasingly relevant for customer success, product, and revenue leaders: What AI can’t tell you about customer decisions and churn.
As organizations race to adopt and understand AI-powered tools for churn analysis and customer intelligence. there is optimism that technology will unlock the secrets behind customer behavior. AI can rapidly analyze vast amounts of data, identify patterns, summarize interactions, and even surface risk signals.
Yet one critical question remains:
Can AI truly explain why customers make the decisions they do?
The answer, at least today, is more complicated than many organizations realize.
Every SaaS company is investing in AI, Ruben pointed out.
AI analyzes support tickets, scores account health, summarizes calls, detects signals, and automates workflows. With so much data available, it's tempting to believe we're finally approaching a world where technology can explain exactly why customers buy, renew, expand, or leave.
But there's a problem: AI can tell you what happened. It still struggles to tell you why. When it comes to customer decisions, that distinction matters.
We Don't Have a Data Problem
Every customer interaction leaves a digital footprint, meaning the modern SaaS organization is drowning in data. It has:
Product usage analytics
Support cases
CRM activity
Conversation intelligence platforms
Health scores
Customer surveys
Renewal forecasts
AI-generated summaries
Yet despite all of this information, many organizations still struggle to answer questions like:
Why did that customer stop engaging?
Why did onboarding stall?
Why did the executive sponsor disappear?
Why did a seemingly healthy account churn?
The Story AI Can't See
The challenge isn't data collection – it's interpretation, and that’s where humans excel. Data points show behavior but don't always explain motivation or meaning. For example, AI can detect that engagement dropped, but it cannot always understand the tension in a boardroom meeting. Or, it can identify a decrease in usage, but it cannot understand the frustration behind a stakeholder's decision to abandon a project.
Those unspoken concerns frequently become the real drivers of customer decisions.
The Hidden Gap Between Signals and Decisions
One of the most consistent findings from IcebergIQ’s customer interviews is that major decisions are rarely impulsive, Natasha said.
Based on churn analysis work with 100+ B2B SaaS companies, IcebergIQ has found that many customers begin considering alternatives 12 to 18 months before a contract is canceled. By the time the renewal discussion happens, the decision-making process may be well underway. Leading up to renewal, competitors are building relationships. Alternative solutions are being evaluated. Internal stakeholders are discussing concerns, and confidence is either growing or deteriorating.
None of these factors may be visible inside a CRM. The customer journey isn't just a sequence of activities, it's an emotional experience – which doesn't always leave a digital trail.
Why Customer Interviews Still Matter
This is where qualitative third-party research becomes indispensable.
At IcebergIQ, our work revolves around structured customer interviews designed to uncover the unbiased, full story behind customer decisions. Our methodology – developed by Alan Armstrong, founder of our predecessor company, Eigenworks – focuses on understanding the customer journey from beginning to end:
What were they trying to achieve?
What expectations did they have?
What worked?
What didn't?
When did things begin to change?
What ultimately influenced their decision?
What could have altered the outcome?
Across thousands of customer interviews, we see consistent patterns: Customers don't often churn just because of product issues. They churn when trust begins to break down.
Sometimes trust is damaged during implementation. and other times it's affected by unmet expectations, communication gaps, or small disappointments that accumulate over time.
The technical issue is rarely the entire story. It’s the emotional impact of the issue that often matters more. That's why organizations that focus exclusively on operational metrics, and don’t have direct conversations with their customers, can miss critical warning signs.
The Future Isn't AI or Humans. It's AI Plus Humans
None of this diminishes the value of AI. In fact, it is becoming increasingly effective at identifying patterns, monitoring signals, automating workflows. and improving efficiency.
These capabilities are incredibly valuable, but they work best when paired with human understanding. AI can help organizations know where to look. Human conversations reveal what actually matters.
The companies that thrive in the next decade won't choose between artificial intelligence and human intelligence – they'll combine both. They'll use AI to detect signals faster, and they'll use customer research to understand the meaning behind those signals.
Best Practices for Implementing Customer Interviews
Customers will open up more freely to a neutral third party like IcebergIQ, Natasha pointed out. But whether you work with a third party or conduct your own internal research, here are best practices for organizations looking to complement their AI investments with deeper customer understanding:
1. Build Executive Alignment Around Transparency
Customer research is most effective when leadership commits to hearing the truth, even when it is uncomfortable. The goal is understanding, not assigning blame.
2. Start with Specific Research Questions
Define the questions you want to explore and the customer segments you want to understand before launching a program.
3. Prioritize Recent Churns or Losses
Interview customers involved in the most recent quarter's wins, losses, renewals, or churns. Fresh experiences yield more detailed and actionable insights.
4. Use Multi-Channel Recruitment
A combination of phone outreach and email outreach typically generates the strongest participation rates.
5. Offer Appropriate Incentives
Providing a gift card or charitable donation can increase participation rates. The amount may vary depending on the person’s role.
6. Integrate Findings Into Your Voice-of-Customer Program
Customer interviews should not be a standalone exercise. The findings should feed product strategy, customer success initiatives, sales enablement efforts, and executive decision-making.
7. Turn Insights Into Action Immediately
Assign owners and next steps as reports are delivered. The value of customer research comes not from collecting insights but from acting on them.
Learn More
To learn more about IcebergIQ's win-loss and churn analysis programs, contact us at info@icebergiq.com or schedule a chat with Natasha Narayan here.