The 98% problem nobody talks about

A tech stack of one

By the time I left Scale AI in 2024, I was managing accounts worth over $100 million. Every morning, I'd open my laptop to 200+ unread messages across email, Slack, and four other chat services. By lunch, it was 400. By end of day, closer to 600.

And not spam. Real requests from real customers on real deals. Security questionnaires. Pricing model updates. Tech specs. Exec intro requests. Contract redlines. Reference calls. Plus the internal fire drill: leadership wants all those conversations added to the updates, in spreadsheet and slideshow format, by EOD for tomorrow's biweekly sync.

Each one mattered. Each one could accelerate or kill a deal (well, not the internal syncs, but that’s a different issue). And each one required me to stop everything, dig through five different systems, craft a response, and hope I didn't miss something critical.

Our tech stack for managing this chaos? Google Sheets.

Seriously.

Not for lack of trying. We had every tool in the book - CRM, RevOps platforms, sales enablement, project management, automation. But when a customer asked for something specific? The answer was never where you needed it. You'd scramble to find it, build it from scratch, then dump it in the CRM where it would slowly rot until someone else needed it months later and discovered it was hopelessly out of date.

Meanwhile, vendors kept pitching us AI tools to "revolutionize" our sales process. No good. Every single one was solving the wrong problem.

The 98% problem no one talks about

Here's the dirty secret of AI in sales: everyone's fighting over the same 2% of the sales cycle.

Walk the exhibit hall at any tech conference and you'll see fifty companies promising to revolutionize your SDR team. Another thirty swear their AI can enrich your leads better than anyone else's AI. They're all solving for the top of the funnel – that first touch, that initial qualification, that opening email.

And sure, if you're selling $50/month subscriptions to consumers or $500 SaaS seats to SMBs, maybe that first touch is most of your battle. Send enough emails, qualify enough leads, and the law of large numbers takes over.

But enterprise sales doesn't work like that.

When you're working a $2 million deal over 18 months, the SDR handoff happens in week one. The enrichment data gets pulled on day two. Then what? You've got 77 more weeks of actual selling to do. Hundreds of stakeholder interactions. Dozens of "small" requests that can make or break your champion's credibility. Technical deep-dives. Security reviews. Procurement negotiations. Executive alignments.

Every one of those moments can either accelerate your deal or kill it. Miss a security questionnaire deadline? Deal slips a quarter. Fumble the technical deep-dive? Your champion loses faith. Send the wrong pricing model to procurement? Welcome to discount hell.

These aren't edge cases. They're the entire job. 86% of deal slippage happens in the middle of the sales cycle, not at the beginning: this is where deals are actually won and lost. And it’s where every sales rep is still working like it's 1995 – copying and pasting between systems, manually tracking follow-ups, praying they didn't miss something critical in the noise.

The entire AI sales tool industry is so obsessed with replacing SDRs that they forgot about the other 98% of the sales cycle. The part where the real work happens. The part where enterprise deals actually get done.

Those are the moments we're building for.

The moments everything changes

At gigue, we call them "flipturns."

Ever watch Olympic swimming? A flipturn is that movement when you tuck underwater right before the wall, rotate, and push off in the opposite direction. Done right, it's almost effortless and sends you flying halfway down the lane. Done wrong, you lose all momentum and pop up early, sputtering for air while your competitors shoot ahead of you.

In enterprise sales, a flipturn happens every time a customer asks for something: a pricing update, a security questionnaire, a reference call, a contract revision. They happen constantly, and they’re the moments that determine deal velocity. Handle them fast and you accelerate toward close. Fumble them and deals die in pipeline purgatory.

Right now, every flipturn requires you to stop what you're doing, context switch between five different tools, hunt for information, craft a response, and hope you didn't miss anything. 

It's like the difference between a kid's first swimming lesson and an Olympic athlete. 

The kid clings to the wall, terrified of going under because they know they'll come up gasping and flailing. 

The Olympian? They've practiced the movement ten thousand times. They trust their technique. They know exactly how much air they need, exactly when to tuck, exactly how hard to push off.

The legacy revenue tool stack forces sales reps to hit the wall each time, manually handling every customer request because they know one wrong move could flood the deal. 

What's worth building

But what if you had that Olympic-level confidence? What if every flipturn was smooth, powerful, and actually shot you forward?

What if your workspace knew what the customer needed, had the context ready, drafted the response in your voice, caught the mistakes before they went out, and queued up the next steps automatically?

Not because some chatbot is pretending to be you. Because you have a real workspace – like developers have IDEs – where everything you need is right there, where AI is ambient instead of intrusive, where you can focus on the human parts of selling while the machines handle the mechanical parts. 

Why you should care

The 98% problem isn't just about sales efficiency. It's about rethinking how humans and AI work together, from the ground up.

The previous generation of sales tools treats salespeople like a problem to be contained. “Hem the AE in.” “Force adherence.” “Correct the language.” “Stop the error midcall.” Whether you’re a new rep trying to learn or the top performer at your company, the system sees you as a cat to be herded.

The current generation of AI tools takes it further: reps are a problem to be removed altogether. "Replace the SDR." "Automate the follow-up." "Eliminate human touchpoints." 

But that's backwards. The human relationship is why enterprise deals close. Trust is why champions stick their neck out, even if the solution isn’t 100% ready. Expertise is why technical evaluations pass, even if that 160-point checklist only has 150 checkmarks. Connections are why sponsors open doors and close contracts.

What we need isn't AI that replaces salespeople. We need AI that makes salespeople superhuman at the moments that matter. The flipturns. The 98% of the sales cycle that actually determines whether deals close or die.

What’s next

Over the next few weeks, I'll be diving deeper into how we're solving this at gigue. How we're building the first real IDE for business work. How we're turning every flipturn from a momentum-killer into a momentum-builder. And how we're helping the best salespeople in the world focus on what they do best: building relationships and closing deals.

But more than that, we're exploring what it means to build software that amplifies human capability rather than replacing it. Because if we get this right for sales, we can get it right for every knowledge worker. Because, as it turns out, lots of other folks are drowning in the 98% problem too.

The future isn't AI or humans. It's both, working together, each doing what they do best.

And it starts with fixing the flipturns.