Findings From a Year of Running Our Sales in Sales AI

A co-founder’s look back at what actually changed once we started operationalizing our own customer intelligence — not the pitch, but what we found in practice.

There was no shortage of information in our sales process before any of this. We had CRM data. We had emails. We had customer meetings. We had call recordings and transcripts. We had notes, tasks and opportunity updates. And by the time we started building Sales AI, we already had AI that could summarize almost all of it.

But a year in, what stands out to me as a sales leader isn’t that we had more information. It’s that we finally had a way to use that information to run the business.

That distinction has only gotten clearer the longer we’ve run our own sales organization on Sales AI. What we’ve come to value most isn’t that it takes notes or summarizes a meeting. It’s that it let us operationalize what we learned from our customers.

I Could Form My Own View of an Opportunity

This was probably the biggest change for me personally over the past year. When I wanted to understand an opportunity, I stopped needing to ask the salesperson for an update. I could look at the information we’d captured across customer meetings, emails and CRM, and just start asking questions.

What did we actually know about this opportunity? What didn’t we know? What were the customer’s priorities? Who had we really engaged with, and at what level? What had changed in the last few weeks? What commitments had the customer made? Where were the risks? Were there contradictions between what we believed internally and what the customer had actually told us?

I didn’t have to be in every customer meeting. I didn’t have to read every email. And, importantly, I didn’t have to ask my salesperson to put together another report for me. I could develop my own perspective from information that was already there and that changed how I could contribute as a manager.

Our Deal Reviews Became Discussions, Not Reporting Sessions

Think back to what our traditional forecast calls used to look like. A manager would ask about an opportunity. The salesperson would explain what happened. We’d discuss the latest meeting, whether the customer was engaged, who we were talking to, and whether the deal was really going to close. A significant part of that conversation was essentially information transfer.

With Sales AI, we started coming into that conversation with our own understanding of the opportunity already formed. We could see the actual engagement, how well we were qualified, the latest customer interactions, which stakeholders we’d engaged with, the level of those relationships, what was agreed, and what still needed to happen.

So instead of asking, “Can you give me an update on this deal?” I found myself starting with something much closer to where the real conversation needed to be:

“I see we have strong engagement with these people, but I’m concerned that we still don’t have access to the economic buyer. Mark seems to have Champion traits — could he connect us?”

That turned out to be a very different kind of conversation. We spent less time reconstructing what happened and more time deciding what to do next. And when it came to forecasting, I could combine the salesperson’s judgment with my own assessment of the evidence behind the opportunity.

We Became More Fact-Based

This was another outcome we didn’t fully appreciate at first. Sales conversations naturally contain a lot of interpretation: “The customer really liked the solution.” “I think we’re in a strong position.” “They want to move quickly.” “We have good access to the decision makers.”

Those judgments still mattered – selling will always require human interpretation. But wherever we’d captured customer communication, through meetings, emails and CRM information, we could go back to the underlying signals. What had the customer actually said? What positive signals did we have, and what negative ones? What commitments were made? What evidence supported our qualification?

It didn’t remove judgment from sales management. It made that judgment better informed, and more fact-based.

The Information Didn’t Stop in Sales AI

This is where the distinction between capturing intelligence and operationalizing it became really important for us. It wasn’t enough for Sales AI to understand that something had changed — that information needed to become part of the sales operation.

If we learned something relevant to an opportunity, Sales AI could identify it and prompt the user to approve the appropriate CRM update. If we discovered a new stakeholder, that became part of our understanding of the account. If a meeting revealed a qualification gap or a risk, we could act on it. If the customer committed to something, we knew whether it happened.

For our Sales Operations team, that had an obvious benefit: the CRM had a much better chance of reflecting what was actually happening in the field, without requiring salespeople to manually document everything they already knew. Sales Operations got better data. Management got better visibility. And salespeople spent less time translating customer conversations into administration.

One of Our Favorite Features Turned Out to Be Tasks

This might sound less exciting than generative AI, but it turned out to be one of the hidden gems. Think about how many actions get created during the lifetime of an opportunity: “We’ll send you the business case.” “I’ll introduce you to our security team.” “Can you provide the technical documentation?” “Let’s schedule a session with the CFO.” “I’ll come back to you next week.”

These commitments used to be scattered across emails, meetings, CRM notes and people’s heads. Sales AI captured them and brought them together through Task AI, so I could look at an opportunity not only from the perspective of what do we know?, but also: what did we agree to do, who owns it, and is it actually happening? That’s operational intelligence. We use it for our internal meetings too — action items committed by sales-supporting colleagues, not just reps, are tracked and documented the same way.

One additional thing we found, looking back over the year: we simply spent less time documenting, organizing meetings, or chasing people to keep customers and internal contacts informed. It made our operation more effective and just as important, more accountable.

This Is Why We Don’t Think the Future Is About Better Note-Taking

We sometimes get asked how Sales AI compares with AI note takers, or the growing number of general AI tools that can analyze meetings, documents and emails. It’s a fair question – the same advances in LLMs that made those products possible are what let us build Sales AI in the first place.

But the more interesting question, a year in, is what happens after AI understands the conversation. A meeting was never an isolated piece of content for us. It was part of an opportunity. It connected to emails, previous meetings, stakeholders, qualification, CRM information, commitments, tasks, pipeline, and ultimately the forecast.

This is why our focus at iSEEit has always been on sales, integration and automation. Looking back, I increasingly think about it in three steps:

Capture → Understand → Operationalize.

Capture what’s happening across the customer relationship. Understand what it means in the context of the opportunity. And then operationalize that intelligence so salespeople, managers and Sales Operations can actually use it in their daily work. A year in, the third step is still where most of the value was created.

The Real Test for AI in Sales

AI could already generate an impressive meeting summary in seconds a year ago. But that was never the test I cared about most. My test has always been what happens on Monday morning.

Did what we learned help the salesperson progress the opportunity? Did it expose a risk we should address? Did it keep our CRM current? Did it remind us about a customer commitment? Did it improve the quality of our deal review? Did it help me make a better judgment about the forecast? And did it do all of this without creating another layer of work for the sales team?

Looking back on the past year, that’s what making customer intelligence operational has meant for us. As a sales leader, I never wanted more information to consume — I wanted the information we already had to help us operate better. And as a co-founder of iSEEit, that’s become one of the principles I care most about as we keep building Sales AI: turn customer interactions into intelligence, and turn that intelligence into action.

See how it works: now.iseeit.com/sales-ai