The Art of Calling Bullshit

Last week, a CEO told me his company spent $2 million on AI. When I asked what decisions it helped him make differently, he went quiet.

Then he admitted: "We're still figuring that out."

He's not alone. Right now, AI conversations are drowning in bullshit from three directions:

• Vendors selling magic

• Teams justifying budgets

• Executives creating their own fiction.

It's time we called out the bullshit. All of it.

The Vendor

AI vendors have perfected the art of the impossible promise. They walk into your boardroom with polished demos, ROI calculators, and case studies that would make a fiction writer jealous. But vendors follow patterns, and once you know them, they're easy to spot.

The "Too Good to Be True" Test

Any vendor promising 80% efficiency gains is lying. Full stop. Real AI implementations deliver 15-30% improvements. Sometimes less. The bigger the promise, the bigger the lie.

Here's what else fails the test: When they can't explain their solution in plain English. I watched a vendor spend 45 minutes explaining their "cognitive decision engine." Turns out it was a rules-based system with some natural language processing bolted on. If they need jargon to make it sound impressive, it's not. And yes, before you get angry with me, these are valid technologies...but just be honest about what is happening in the 'black box.'

Perfect demo data is the third tell. Every vendor demo runs on cleaned, curated data that looks nothing like your messy reality. Ask to see it work on a sample of your data. Watch them squirm.

The Hidden Cost Reality

That $500K license? It's maybe 20% of your real cost. The vendor won't mention the six-month integration that turns into 18 months. Or the data scientists you'll need to hire. Or the three systems you'll need to rebuild because they don't play nice with AI.

One manufacturing client learned this the hard way. The vendor quoted $300K. The final cost? $1.8 million, including integration, training, consultants to fix what the first consultants broke, and the temporary staff to keep things running during the "seamless transition."

Red Flag Phrases

Learn these phrases. When you hear them, your BS detector should start screaming:

• "No technical expertise required" means you'll need three PhDs to keep it running.

• "Seamless integration" means at least six months, if you're lucky.

• "Real-time insights" means dashboards that update every 24 hours.

• "Revolutionary approach" means they added AI to something that already existed.

The Internal Team

Your team creates a special flavor of BS. They're not lying exactly; they're just caught between reality and what they think you want to hear.

Warning Signs from Your Team

When your team needs 47 slides to explain the business case, there is no business case. Real value explains itself in five slides or fewer. The other 40+ are camouflage.

Watch for shifting metrics:

• Month one: "We'll reduce processing time by 50%."

• Month three: "We're seeing improved user satisfaction."

• Month six: "Adoption rates are climbing."

When the goalposts keep moving, you're funding a science project, not a business initiative...and that might be ok. Just be aware that it's a science project.

The worst tell? When "everyone else is doing it" becomes the primary justification. I heard this from a CFO whose team pushed for an AI-powered forecasting tool. When pressed, they admitted their current Excel models worked fine. They just felt left behind.

Questions That Cut Through Internal BS

Three questions expose 90% of internal AI BS:

1. "What specific decision will this help us make differently?" If they can't name it, stop there.

2. "Show me where this failed at another company." If they've only researched success stories, they haven't done their homework.

3. "What are we NOT going to do if we do this?" Every yes needs a no. If they can't tell you what they'll stop doing, they're just adding complexity.

The Executive (ahem…you)

You may not want to hear this, but executives generate more AI BS than vendors and internal teams combined. You create pressure that forces bad decisions, then wonder why projects fail.

The Bullshit You're Creating

It starts at conferences. You hear a keynote about AI transformation. You see competitors announcing AI initiatives. You return to the office and declare: "We need to be AI-first by year-end." No problem definition. No success criteria. Just FOMO dressed up as strategy.

The "vision without operations" problem is worse.

"AI will transform our business model!"

How?

"It will make us more efficient!"

How specifically?

"That's what we need to figure out!"

This isn't leadership. It's pure theater.

Then comes budget theater. You cut headcount to "invest in AI capabilities." Then hire consultants at 3x the cost to tell you what to do with the technology. Then, you hire more consultants to implement the recommendations of the first consultants.

Meanwhile, the people who knew your business are gone either because you cut them or they left because you didn’t value their input.

The Responsibility Duck

Watch how AI initiatives flow through organizations. The CEO says, "Make it happen." The COO assigns it to "whoever handles technology." The CTO says it's "really a business transformation issue." Round and round it goes.

"The tech team will handle it" might be the most expensive sentence in business. AI isn't a tech problem. It's a business problem that uses technology. When you make it IT's responsibility, you guarantee a technically perfect solution to the wrong problem (usually).

"Report our AI progress to the board" creates another special kind of BS. What progress? The chatbot that handles 5% of customer inquiries? The pilot that 12 people use?

A Framework That Works

After watching hundreds of AI initiatives crash and burn (and a few succeed), patterns emerge.

Here's an approach that works.

For Vendor Pitches

• Demand failure stories, not success stories. Any vendor can show wins. Make them explain their disasters. How did implementations fail? What went wrong? What did they learn? If they claim they've never failed, end the meeting.

• Calculate the total cost upfront. Take their number and multiply by 5. That's your starting point. Include integration, training, maintenance, overhead, and organizational disruption. If it still makes sense at 5x, proceed.

• Get specific about decisions. Not "better insights" or "improved efficiency." Which exact decisions will be made differently? By whom? How often? If they can't map it to specific decision points, they're selling you expensive dashboard decoration.

For Internal Proposals

• Require plain English. If your team can't explain it without jargon, they don't understand it. "We'll use machine learning to optimize customer journey touchpoints" means nothing. "We'll predict which customers will leave and call them first" means something.

• Set kill criteria upfront. Before starting, define what failure looks like. Adoption below 50% in 90 days? Kill it. Accuracy below human baseline? Kill it. Cost exceeding 2x projection? Kill it. Make the criteria clear and stick to them.

• Define success without using "transform," "revolutionize," or "disrupt." Real success is boring. "Reduce invoice processing from 4 days to 1 day." "Cut customer churn by 10%." "Eliminate 3 manual reports." Boring. Measurable. Valuable.

For Yourself

• Stop FOMO-driven mandates. Your competitors' press releases aren't strategy documents. Half their "AI initiatives" are probably failing too. Focus on your actual problems, not their marketing.

• Own the strategy or don't start. AI initiatives without executive ownership fail 100% of the time. Not 90%. Not 95%. All of them. If you're not willing to own it personally, save everyone the trouble.

• Measure outcomes, not activity. "We have 5 AI pilots running" isn't a metric. "We reduced customer acquisition cost by 12%" is. Count decisions improved, costs reduced, and problems solved. Everything else is theater.

The Bottom Line

AI bullshit is expensive. But pretending it doesn’t exist is even more expensive.

I've watched companies burn through millions of dollars chasing technology and AI dreams while their real problems sat untouched. The vendors got rich. The consultants got richer. The problems got worse.

The fix isn't complicated. It’s a simple fix. Stop lying to your board, your team, and yourself.

Most AI initiatives are solutions looking for problems. Most demos are just theater. Most ROI projections are fiction. And most executives know it, but fund it anyway.

The companies winning with AI share one trait: they call bullshit early and often. On vendors. On their teams. On themselves. They'd rather kill a bad idea in month one than explain a disaster in year two.

Now, it’s your move. Keep funding the fiction, or start calling out the bullshit.

This is the kind of strategic clarity I help executives achieve. If you're ready to cut through the AI noise and focus on what drives results, let's have a conversation. You can find me at ericbrown.com or connect on LinkedIn.

If you found this post helpful, consider sharing it with another executive grappling with AI, technology, and data. If you want to explore AI and other Technology strategies, grab some time on my calendar, and let's chat.

Newsletter Recommendations