Ask most revenue operations leaders how they know why customers buy, and the answer arrives in one of three shapes. A pick list value in the CRM. A hunch from the rep who was in the room. A tracker someone built in a call recording tool eighteen months ago and never revisited. It works, roughly, right up until you have to make a real decision with it.
Chad Boersma spent more than ten years in revenue operations before joining Airspeed as GTM Agentic Architect Lead. He was an Airspeed customer before he was a colleague, and he lived inside that gap for most of that decade.
“For a long time you had to use sales input to gather all of this data. We had to use a pick list value in the CRM. We had to use anecdotal evidence from sales. We really weren’t understanding the voice of the customer.”
Chad Boersma, GTM Agentic Architect Lead, Airspeed
The shortage was never conversations. Teams were recording calls, running discovery, sitting through renewals and QBRs. The shortage was time. All of that qualitative signal, the actual language customers used about their problems, sat locked inside recordings nobody could get through.
Why quantitative analysis alone falls short
Revenue teams have become very good at counting things. Pipeline created, win rates, cycle length, activity volume. Those numbers tell you what is happening. They rarely tell you why.
“You try to use quantitative analysis to understand why customers were purchasing from us, why they would not, what is the messaging that is winning with customers,” Chad says. The intent was right. The tooling forced a compromise, because turning a two hundred call quarter into structured insight meant either flattening it into a dropdown or spending days listening back.
Think about what a pick list actually asks a rep to do. It asks them to take a rich, forty-five minute conversation, full of hesitation, competing priorities, and the specific words a buyer used, and reduce it to one of eight predefined reasons. Closed lost: price. Closed lost: timing. Closed lost: no decision.
The moment you do that, the texture is gone. You can count how often price was selected. You cannot see that in half those deals price was standing in for a weak business case, or that a competitor kept surfacing in the same breath.
Where keyword trackers hit a wall
Chad’s team had used a conversation intelligence tool for exactly this reason, and still hit a wall. “We’d use call recordings, which would take a long time to review those things, or set up trackers and make sure that the analysis that was coming out of that tool was effective and useful.”
Trackers are keyword spotters. They find the words you already thought to look for. They cannot tell you about the objection you did not anticipate, or the phrase three prospects used last week that signals a shift in the market.
They also carry a maintenance cost of their own. Someone has to define them, tune them, and keep checking that the output still means something as your product and your market move. The analysis is only ever as good as the list of things you remembered to watch.
That is the quiet trap. You buy a tool to understand your customers, and you end up spending your time managing the tool instead of learning from the conversations. It is one of six places conversation intelligence loses the insight between the call and the decision.
What changes with AI-native analysis
The step change is moving from keyword matching to concept understanding. An AI-native platform reads the whole conversation, understands your business and your market, and produces qualitative analysis you can take into a strategy meeting.
“We wanted something that was AI native that could really look through our conversations with our customers, understand concepts, understand our business, understand our market, and develop qualitative analysis that was beneficial and was used in strategic decision making.”
Chad Boersma, GTM Agentic Architect Lead, Airspeed
Understanding concepts rather than words is the whole difference. A buyer does not have to say data residency for the system to register that a compliance concern about where information is stored just came up. It reads meaning, the way a sharp analyst would if that analyst had listened to every call your team ran this quarter and remembered all of it.
What each go-to-market team gets
In practice, different parts of your motion get answers pitched at their level, from the same source.
Product sees which features are winning and losing, and where. Chad describes running agents “to understand what are the product features that are most beneficial, where are we winning and where are we losing on feature sets.” That feeds the roadmap with evidence instead of the loudest opinion in the room.
Marketing sees which messages land, and which metrics and outcomes customers actually care about, so campaigns speak to real buying criteria rather than internal assumptions.
Sales sees the discovery and demo talking points that correlate with closed won deals, which is a far better coaching input than a manager’s memory of a good call.
All of it comes from real customer conversations, not from three different recollections of them. Because the capture is automatic, none of it costs a rep an hour of CRM admin either.
How voice of the customer becomes a strategic input
This is the part that matters. Voice of the customer stops being a survey you run once a year and becomes a live input into how you sell, build, and message.
The teams getting the most from this treat their call library as a source of truth rather than an archive. When a competitor changes their pricing, you can see how prospects react to it that week. When a new objection appears, you catch it in week one, not in next quarter’s board deck. When a message starts to lose its edge, the calls tell you before the pipeline does.
That only works if the analysis runs on everything, and if you can ask it questions nobody set up in advance. That second part is what separates AI-native conversation intelligence from a keyword tool. A sample of the calls someone had time to review is still an anecdote, just a better dressed one.
Where to start
Pick the question your leadership team argues about most. Why we lose to one particular competitor. Which objection is slowing the mid-market segment. Which feature gap keeps showing up in stalled deals.
Then ask it of your actual conversations rather than your CRM fields, and see how close the two answers are. The gap between them is the size of the problem, and call intelligence is where you close it.
If your voice of the customer program still runs on pick lists and gut feel, the constraint is not your team. It is tooling that forced them to compress rich conversations into thin data. The conversations are already happening. The only question is whether you can learn from all of them, or just the handful someone had time to review.
See how Airspeed turns customer conversations into qualitative analysis your whole go-to-market team can use. Book a demo.
Frequently asked questions
What is voice of the customer analysis?
Voice of the customer analysis is the practice of turning what customers actually say into insight a revenue team can act on. Traditionally it ran on surveys, CRM pick lists, and rep anecdotes, which compress a long conversation into a single dropdown value. AI call analysis reads the full conversation instead, so the insight reflects the customer's own language rather than a rep's summary of it.
Why is quantitative analysis not enough to understand customers?
Pipeline created, win rate, cycle length, and activity volume tell you what happened. They rarely tell you why. A closed lost reason of price cannot show you that price was standing in for a weak business case in half those deals, or that the same competitor kept surfacing alongside it. The why lives in the conversation, not the field.
What is the difference between keyword trackers and AI call analysis?
Keyword trackers find the words you thought to look for. They cannot surface the objection you did not anticipate, and they carry a maintenance cost because someone has to define, tune, and re-check them as the product and market move. AI call analysis reads for concepts, so a compliance worry about where data is stored registers even when nobody says the phrase data residency.
How do go-to-market teams use voice of the customer analysis?
Product teams see which features win and lose, and where. Marketing sees which messages land and which outcomes buyers actually care about. Sales sees the discovery and demo talking points that correlate with closed won deals. All three read from the same set of real customer conversations rather than three different recollections of them.
How often should voice of the customer analysis be refreshed?
Continuously. An annual survey tells you what customers thought two quarters ago. When analysis runs against every call, a new objection shows up in week one, a competitor's pricing change shows up in how prospects react to it, and a message losing its edge shows up before the pipeline reports it.
