Last month I got on a call with the CEO of a music publishing company.
Smart guy. Built a real business. He told me he'd been spending 3 to 4 hours every weekday inside Google Ads. Manually adjusting bids. Copying campaigns. Pulling reports.
Then he said the line that's stuck with me:
"This work has nothing to do with actually growing the business."
He'd known that for over a year. He hadn't fixed it. He's not stupid. He's not lazy. He's just one of the 95%.
That number isn't mine. It's from a report MIT published called State of AI in Business 2025. They studied how companies are actually using generative AI inside their operations. The finding made headlines across Fortune, Bloomberg, and the Wall Street Journal: 95% of corporate AI pilots are delivering zero measurable return. Only 5% are pulling real value out of it.
(Source: MIT NANDA, State of AI in Business 2025. Fortune's writeup is here)
I sat on 50 inquiry calls with business owners over the last six months, and the MIT report reads like a transcript of my notebook.
A 40-location dermatology group missing roughly 100 phone calls a day. They've known for months. Their conversion rate on calls they DO answer is 65%. Every unanswered call walks next door.
A dementia care company where half of inbound calls go to voicemail. Live calls convert at 90%. Voicemail callbacks convert at under 5%. They are quite literally losing the families that needed them most.
A wholesale distributor with 60 salespeople trying to cover 15,000 customers. Their existing chatbot runs at a 2% success rate. They've known for over a year.
Same pattern. Different logo.
The MIT report is clear on why most companies fall into the 95%: it isn't the model. It's the wiring. The companies in the winning 5% picked the workflow that was actually bleeding them and built AI around it. The other 95% bought tools and never connected them to anything that mattered.
McKinsey's State of AI report puts harder numbers on the gap. The 5% are growing revenue close to 2x faster than everyone else. Margins are widening 8 to 10 points. And the gap is compounding every quarter, not closing.
Here's the part most business owners aren't ready to hear: even the people who think they're using AI well are typically using it at about 10% of what it can actually do.
So I shot the most complete Claude Code walkthrough I've ever put out. 6 levels. 35 concepts. Every feature that shipped in the last 90 days the older tutorials skipped: managed agents, routines, sub agents running in parallel, work trees, hooks, Opus 4.7. From your first prompt to Claude running autonomously in the cloud on a schedule while you're asleep.
If you've felt a step behind on AI, this is the one to watch.
-> Watch the Walkthrough here.
If you'd rather have us build it into your operation for you, customized to your tools and your team trained on it, we've done this for 40+ companies and driven $5M+ in bottom-line revenue across them. Book a free call and we'll map out exactly what it looks like for your business.
-- Nick
P.S. The CEO of the music publishing company figured out how to automate his Google Ads work in about two weeks. He got his mornings back. The diagnosis is usually free. Execution is the whole game.