The signal the market just sent
When the company that holds more customer data than anyone in the world goes out and acquires a conversation intelligence company, it’s worth pausing to understand what that move is saying.
Salesforce — owner of Salesforce CRM, Service Cloud, and a significant portion of the systems where commercial relationships are managed — agreed to acquire Cresta, a platform focused on voice agents, real-time assistance, and conversation intelligence. The move matters not because of the technology itself, but because of what it acknowledges: Salesforce had the customer context and the systems. What it was missing was visibility into what actually happens inside each conversation.
That gap — between having customer data and understanding what takes place in interactions — isn’t a problem unique to the world’s largest CRM. It’s the same gap most commercial teams have today, without having diagnosed it yet.
Customer data vs. conversation intelligence: they’re not the same thing
It sounds like a subtle distinction. It isn’t.
Having customer data means knowing a client’s history, the deal stage, the segment, the contacts involved, prior logged interactions. All of this lives in the CRM — if someone put it there.
Understanding what happened in a conversation is something else entirely: what the customer said, what objections came up, what the real tone of the exchange was, what commitment was made at the end of the call, and whether the rep actually addressed the pricing concern that surfaced twelve minutes in.
History, segment, deal stage, contacts. Context. What the CRM captures when someone updates it.
What was said, how it was said, what objections surfaced, what real commitment was made. Evidence. What almost never makes it into the CRM automatically.
One is context. The other is evidence. And deal outcomes don’t come from context — they come from what happens in conversations.
The real gap: what happens on the call never makes it into the CRM
Here’s the structural problem. Everything that happened in a conversation only reaches the CRM if the rep decides to write it down — and writes it accurately, completely, before moving on to the next call.
What that process produces in practice:
- Key objections live in the rep’s head, not in the system.
- Deal risk becomes visible when the deal is already lost.
- Forecast reviews are based on the rep’s interpretation, not on what was actually said.
- Managers can’t evaluate what they can’t see.
- When the rep leaves, that information leaves with them.
The data in the CRM reflects what the rep decided to log. Not what happened.
Why context isn’t enough without the conversation
There’s a prevailing belief in many teams: if you have enough customer data, you can predict deal outcomes. CRM hygiene advocates argue exactly this — more fields, more updates, better visibility.
The problem is that deals aren’t won or lost in a CRM field. They’re won or lost in conversations. And conversations, by default, are invisible.
A team can have an impeccably updated CRM — every stage moved on time, every field complete — and still have no idea why deals are stalling. Because what stalls a deal almost always happens in a conversation, not in a form.
Context tells you who the customer is. The conversation tells you what’s actually happening with that customer. Without the second, the first only helps you sort the list.
The three questions your CRM can’t answer
A quick test. For any deal that’s been active for more than 30 days, answer these three questions using only the data currently in your CRM:
- What was the exact objection the customer used on the last call, in their own words?
- What did the rep say about pricing, and what was the customer’s reaction?
- What specific next-step commitment was established at the end of the last conversation?
In most cases, the CRM has the rep’s opinion about those three things — if it has anything at all. It doesn’t have what actually happened. That gap between “what the rep thinks happened” and “what happened” is exactly where the end-of-quarter surprises pile up.
What changes when conversations become data
When every call is captured, structured, and made searchable — not just the ones someone decided to review this week — forecasting stops being an exercise in optimism. It becomes anchored in evidence.
The pattern that emerges in teams that close this gap is consistent: managers stop relying on what reps tell them and start working from what actually happened. Pipeline reviews get faster because the information already exists. Coaching becomes specific because there’s a real record to work from. Forecast accuracy improves because the signals were always there — they just weren’t being captured.
The move Salesforce made with Cresta is, at its core, the recognition of this: customer context alone isn’t enough to make a commercial or service team effective. Visibility into what happens inside conversations was the missing piece. And it’s what’s missing in most teams that today review their pipeline blindfolded.
Performy runs this automatically over 100% of calls — not just the ones someone decided to review this week. Every conversation is captured, structured, and analyzed: objections, commitments, risk signals. What previously lived only in the rep’s memory becomes a searchable record available to the entire team, in real time.