I’m going to tell you something that might sound crazy. For the first two years of my last startup, I banned our sales team from using a CRM. No Salesforce, no HubSpot, nothing. People thought I was insane. My co-founder staged a mini-intervention. But I held firm. Why? Because most CRMs are where good sales processes go to die.
They become a data-entry black hole, a glorified rolodex that reps spend more time feeding than using. Founders, especially first-timers, get obsessed with tracking everything. They think more data equals more control, which equals more revenue. It’s a trap. A costly, time-sucking trap.
We didn’t 3x our pipeline by adding more reps or by tracking more fields in a database. We did it by fundamentally rethinking the relationship between sales and data. We did it by finally embracing a new breed of AI-native CRMs, but not in the way you’d think. Here’s the story of how we got it so wrong before we got it right.
The Glorified Spreadsheet Era
Look, I get it. The urge to implement a CRM comes from a good place. You’ve got a handful of early customers, things are getting messy in spreadsheets, and you’re terrified of dropping the ball. So you do what every business school case study tells you to do: you get a CRM. At my first company, MovieLaLa, we fell right into this. We picked a big-name CRM, spent a month customizing it, and forced our tiny sales team to log every single interaction.
What a disaster.
Our reps, who were brilliant at building relationships, were suddenly spending a third of their day on data entry. Their pipelines looked full, but nothing was moving. The data was a mess of inconsistent entries and subjective notes. I’d look at a report and have no real idea what was going on. Is this deal really at 70% probability, or did the rep just feel optimistic that day? It was garbage in, garbage out.
The fundamental problem is that traditional CRMs were built for managers, not for sellers. They are reporting tools, designed to give executives a false sense of security. They don’t help the person on the front lines close a deal faster. They don’t help them write a better follow-up email. They just create more work.
I remember one specific moment of clarity. I was sitting with our top rep, a guy who could sell ice to a penguin. He had his head in his hands, staring at a Salesforce screen. He told me, “Sahin, I spent two hours last night updating my pipeline. I could have spent that time talking to three new prospects.” That’s when I knew we had it all backward. We were optimizing for tracking, not for selling. That’s when I pulled the plug. We went back to a shared document and a simple weekly pipeline meeting. It wasn’t elegant, but it worked. It forced us to talk to each other, to actually understand the deals, not just the data points.
The AI CRM Headfake
Years later, when I was building RemoteTeam, the pressure to get a CRM started much earlier. We had real revenue, a growing team, and VCs who wanted to see hockey-stick projection charts. The term “AI CRM” was just starting to get hot. Every vendor was slapping an “AI” label on their product, promising it would magically solve all our problems. It would score our leads! It would write our emails! It would tell us which deals would close!
I was skeptical. I’d been burned before. My gut told me this was just the same old trap with a new coat of paint. And for the most part, I was right. The first wave of so-called AI CRMs were just traditional CRMs with a few half-baked features bolted on. The “AI” was mostly just a bunch of if-then statements. It was still a system designed for managers, still focused on tracking over selling.
We tried a couple of them. They were… fine. The lead scoring was a black box. It would tell us a lead was “hot” but couldn’t explain why. The activity logging was still a manual chore. We were paying a premium for a slightly smarter spreadsheet. The team hated it, and I could see the same old patterns emerging – reps spending more time feeding the machine than talking to humans.
I was about to give up and go back to my trusty Google Doc when we stumbled upon a completely different approach. It came from a small startup, one of the many I’ve been fortunate to angel invest in. They weren’t building a better CRM. They were building a system to replace the need for a CRM. It was a subtle but profound difference. Their philosophy wasn’t about tracking sales; it was about accelerating them. This is where everything changed.
The Counterintuitive Framework That Actually Worked
What we built wasn't a single tool, but a system of tools and a new philosophy. It was less about a central database and more about a decentralized, intelligent layer that helped our reps at every stage of the sales cycle. We stopped thinking about “CRM” and started thinking about “Revenue Acceleration.”
The core idea was simple: What if we used AI to help our reps sell, instead of just using it to track what they sold?
This led us to a framework that was completely counterintuitive to everything I’d learned about sales management. It focused on empowering reps, not monitoring them. Here’s what it looked like:
1. Automate the Mundane, Not the Relationship
The first thing we did was identify every single task that wasn't directly related to talking to a customer. Updating the pipeline, logging calls, scheduling follow-ups, writing boilerplate email copy. We automated it all. We used a combination of tools like Gong to record and transcribe calls automatically, and then used custom scripts to pull key information and update a central dashboard. The rep didn’t have to do a thing. Their only job was to have great conversations.
2. Surface the Signal from the Noise
Instead of a black-box lead score, we built our own “deal scoring AI” model. But here’s the key: it was transparent. It wasn’t just a number; it was a story. The system would analyze call transcripts, email sentiment, and engagement data (like how many people from their team viewed our proposal) and then present a simple, human-readable summary. For example: “Warning: The prospect mentioned ‘budget constraints’ three times in the last call. Suggest sending the case study on ROI.” This was a breakthrough. The AI wasn't telling the rep what to do; it was giving them intelligence they could act on. For more on this, you can read my post on the future of deal scoring AI.
3. Instrument the Conversation
This was probably the most important piece. We started treating every sales conversation like a product manager treats a user session. We used AI to analyze everything. What questions were leading to the best discussions? What phrases were causing prospects to disengage? We discovered, for instance, that when we used the phrase “cost savings,” our close rate went down. But when we used “revenue generation,” it went up. We created a library of these insights and fed them back to the team in real-time. The AI would literally pop up a suggestion during a live call: “Try asking about their growth targets instead of their budget.”
4. Shrink the Cycle with Intelligence
Our sales cycle was way too long. The AI helped us figure out why. By analyzing thousands of interactions, we found a huge bottleneck: the handoff from the initial discovery call to the technical demo. It was taking an average of 9 days. The system flagged this, and we realized we weren’t setting a clear next step on the initial call. We implemented a simple change: every discovery call had to end with a scheduled demo on the calendar. The AI even automated the scheduling. That one change cut our sales cycle by 20%.
The Result: More Revenue, Less Management
By flipping the model—by using AI to empower reps instead of monitoring them—we achieved something I didn’t think was possible. We 3x’d our qualified pipeline in six months without adding a single new sales rep. More importantly, our reps were happier. They were spending their days selling, not doing data entry. They felt like owners, not cogs in a machine.
We eventually found an “AI CRM” that fit our philosophy, one that was built around this idea of revenue acceleration. But the tool wasn't the solution. The mindset shift was the solution. You can read more about our approach to building an outbound AI strategy that complements this.
So if you’re a founder struggling with your sales process, I urge you to stop looking for a better CRM. Stop trying to track everything. Instead, ask yourself a different question: How can I use technology to make my sales team smarter, faster, and more effective? The answer probably doesn’t look like a traditional CRM at all.
It looks like a system that gets out of the way. It looks like intelligence, not just data. It looks like trust, not just tracking. And it results in a hell of a lot more revenue.
Frequently Asked Questions
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.