August 28, 2025
Hot Take: Companies Are Buying ChatGPT When They Need Excel

Chris Sterry

Chris Sterry, Co-founder at HelixML, shares his perspective on the “that’s how we’ve always done it” mentality. With over 15 years building and scaling GTM operations across multiple companies—from sales engineer to VP of Operations to Product to founding roles—Chris has been through several acquisitions at Uiflow, Data Republic, and G5, giving him a front-row seat to how technology implementations actually play out in practice. As both a GTM operator and now founder, he has unique insight into the disconnect between what enterprises think they need from AI and what actually drives business outcomes.
🌶️ Hot Take: Companies Are Buying ChatGPT When They Need Excel
My Contrarian Take
Enterprise AI isn't about buying more tools—it's about killing your existing workflow complexity. I know that sounds backwards, but hear me out.
Every executive I talk to is racing to sprinkle AI fairy dust across their organization, burning millions on shiny generative AI platforms while completely missing what's actually happening. Most enterprise AI implementations? They're just elaborate exercises in corporate self-delusion.
Companies keep asking for custom AI integrations when what they actually need is someone to automate their expense reports.
Why They’re Missing the Point
I've sat through more enterprise AI deals than I care to count, and 90% of these pilots fail. Not because the technology sucks, but because companies are trying to force generative AI into their existing, broken processes instead of just... starting over.
It's like trying to make a horse-drawn carriage faster by strapping a tiny motor to it. You're solving the wrong problem.
The pattern drives me nuts. Executives see a ChatGPT demo, get stars in their eyes, then tell their teams to "implement AI across the organization." What they get back is expensive prompt engineering projects that feel innovative but don't actually change how work gets done. Meanwhile, the boring AI that could automate their invoice processing sits there ignored because it's not sexy enough.
From a sales perspective, we're all selling the wrong thing. Customers come to us asking for generative AI because that's what they've heard about. But what they actually need? Basic workflow automation that just happens to use AI under the hood.
The companies that are actually winning with AI do one thing differently: they eliminate entire workflow stages instead of making them slightly better. They're not optimizing their existing mess—they're replacing it completely.
Take contract review. Most companies want AI tools to help their legal teams review contracts faster. But some companies are eliminating the legal bottleneck entirely for standard agreements. They're using AI to automatically negotiate, redline, and approve 80% of vendor contracts without any human touching them. What used to take 6 weeks and kill deals now happens in 24 hours.
That's not incremental improvement. That's elimination of the thing sales teams complain about most.
But selling this is way harder. It's easier to pitch "we'll help your lawyers review contracts 40% faster" than "we're going to eliminate your legal review process for most contracts." One's a software purchase. The other requires convincing general counsel to fundamentally rethink risk management.
One Thing That Actually Works
Founding an AI company, I have the luxury to try new things. After watching too many demos crash and burn, I started looking at my most repetitive workflows, then I challenged myself to eliminate 30% of them.
Not optimize. Not improve. Eliminate.
I picked the stuff that made me want to bang my head against the wall every week. Weekly pipeline reviews where we'd spend 2 hours going through the same spreadsheet asking the same questions. Monthly customer health check meetings that were just status updates in disguise. Those painful quarterly business reviews where everyone pretended the deck was actually useful.
What I found was wild—most of these workflows existed because "that's how we've always done it," not because they actually drove results. So I started experimenting with AI to just... make them disappear.
The weekly pipeline review? Now it's an AI-generated report that hits my inbox every Monday morning with the insights that actually matter. The customer health checks? Automated alerts when something needs attention, plus a monthly dashboard. Those QBRs? The AI pulls the real insights and I just focus on the strategic conversations.
I got 8 hours of my week back. I can spend on deals, on product strategy, on actually building the company.
Then I started asking prospects the same question I asked myself: "What's the one workflow in your organization that, if it just disappeared tomorrow, would make everyone's life dramatically better?"
The responses are incredible. One CFO looked me dead in the eye and said, "If our month-end close process just vanished and the numbers appeared automatically, I'd get 40 hours of my life back every month." That's when you know you're having the right conversation.
People stop thinking "how can AI help my team work faster" and start thinking "what if my team didn't have to do this at all?" The deals that come from these conversations are 10x bigger than the ones that start with feature demos.