TL;DR
A company spent $2M on a sales platform that no one uses. Most teams make the same mistake by skipping small tests and going all in too soon. The smart play is to start small, see what works, and scale only after you have proof.
A CEO just told me his company spent $2M on an AI-powered sales platform. Six months later, it’s sitting unused while his reps stick to their old CRM.
“But everyone’s using AI for GTM now,” he said. “We had to make the investment.”
This is backwards.
While Goldman Sachs predicts AI investment will hit $200 billion globally by 2025, MIT just dropped a truth bomb: 95% of enterprise AI pilots are failing to deliver measurable impact on P&L. Companies are firing massive AI cannonballs without ever testing a single bullet.
Jim Collins called this decades ago. His “bullets then cannonballs” strategy isn’t just smart business, it’s survival in the AI era. Fire low-cost, low-risk bullets to gather data. Once you’ve calibrated your aim, fire the resource-intensive cannonball.
But here’s what’s happening instead: Companies are allocating up to 20% of their tech budget to AI, while purchasing AI tools from specialized vendors succeeds 67% of the time, but internal builds succeed only one-third as often.
Translation: Everyone’s building their own AI Death Star instead of testing with a slingshot first.
The winners? They’re doing the opposite. Startups that dedicate over 50% of their GTM tech stack to AI are seeing 37% reduction in customer acquisition costs – but they got there through rapid experimentation, not massive upfront bets.
Here’s what bullets vs cannonballs looks like in AI GTM:
Bullets: Test ChatGPT for email personalization ($20/month). Try one AI prospecting tool on 100 leads. Pilot an AI chatbot on one landing page.
Gather data: Which AI tool actually improved conversion rates? What tasks does your team avoid vs embrace? Where did AI create measurable ROI vs busy work?
Cannonballs: Only after you’ve proven what works, integrate AI across your entire GTM stack. Build custom solutions. Train your whole team.
The companies firing AI cannonballs in 2025 without testing bullets first aren’t being aggressive, they’re being reckless. Early hypergrowth alone means less now than ever before, and competitive intensity is at an all-time high, with promising areas attracting 2- 3x the rivals of years past.
The shift: Stop trying to win the AI arms race with one big bet. Start with small, fast experiments that prove ROI before you scale. In a world where startups using AI-driven GTM strategies reach product-market fit 2.5x faster, speed beats size every time.
Your next move: Pick one GTM process this week. Test one AI tool for 30 days. Measure the impact. If it works, expand. If it doesn’t, try something else. Fire bullets until you find your target, then bring out the big guns.
Hit reply and tell me: What’s the biggest AI investment your company made that’s sitting unused?
