That’s going to end well… Why does everyone fret about trying to accommodate this aberrant behavior? And yes, I don’t care if it’s 53%, it’s still aberrant because it’s unsatisfactory for good business. Instead set the specific goal of hiring only 2nd & 3rd standard deviation talent. Leave the rest to your competitors….or the co-interviewing parents.

Confidently Wrong! AI is great at data analysis but NOT data engineering xitter.cf/i/broadcasts/1pKkOXkRl…

The risk isn't that AI gives you a wrong answer. It's that it gives you a wrong answer you believe. Great at analysis. Not at data engineering. Most teams ask it to do both.

If the thesis needs another $2M of EBITDA, how many net new sales activities does that take, and can your current team actually produce them? At one portco: 6,000 to 8,000+ activities per rep per year, depending on role and segment. #SalesFunnel #PrivateEquity

Your AI hands you a pipeline number for the board deck. Fast, clean, nice chart. Nothing about it tells you whether it's the truth or a well-written guess.

Anthropic's data team: without structured context, Claude's analytics accuracy didn't exceed 21%. With skills on a governed data foundation, above 95%. Same model. The gap is data engineering, not the AI.

A portco grew revenue 3 straight years. Great waterfall slide. But the growth came from a market with nowhere else to send its money, not the sales team. Buy that at face value and you didn't buy a swimmer. You bought the tide. #PrivateEquity #SalesLeadership

Attackers worked out that shutting down a production line pays more than locking up email or ERP. Production goes to zero. Dragos: ransomware vs. industrial orgs up 87% in 2024, manufacturing hit hardest. Brian C. Reed of Anaconda on the IGI podcast. #OTSecurity

Either your current team can produce enough activity to hit the number, or they can't. Coaching and optimistic forecasts don't change that. It's arithmetic, not motivation. #SalesLeadership #PrivateEquity

A CEO to a seller: “Stop chasing me. Find the owner of the problem in the management tier.” Going high by default isn’t a strategy. Find who actually owns the pain. Brian C. Reed on the Industrial Growth Institute podcast. #B2BSales

Your AI analysis sounds right. That's the problem. More leaders are pointing Claude, ChatGPT and Copilot at their CRM, ERP and marketing data and asking it to explain the business. The answers come back fast, polished and confident. Too often they're also wrong, and nothing in the answer tells you so. Anthropic, the company that makes Claude, tested this on its own internal analytics. Without structured context, Claude answered accurately no more than 21% of the time. With a governed data foundation and skills built on top of it, the same model topped 95%. Their conclusion: analytics accuracy is "a context and verification problem, not a code generation issue." Put simply: AI is great at data analysis. It's really bad at data engineering. And most people using it are quietly asking it to do both. Join Ed Marsh (Ed Marsh Consulting) and Nikos Ntirlis (Databox) for a straight-talk conversation about why "just connect your AI to your data" is dangerous advice, and what to do instead. What we'll cover: 1. Where the data comes from. Why an AI connection to your CRM often returns only part of your data, and why neither you nor the model can tell. 2. Definitions. How "revenue," "pipeline" or "win rate" can mean three different things in three systems, and why the same question can return two different numbers. 3. The context layer. What the model needs to know about your business to map a question to the right data, and why newer models won't figure that out on their own. 4. Why "just connect it" is dangerous. Silent partial data, invented definitions, APIs that change constantly, token bills that pile up, and wrong answers delivered in the same confident tone as right ones. We'll save 15 minutes for your questions, including the one everyone asks: "Why don't we just build this ourselves?" Who should attend: - Private Equity Partners who have been searching for reliable, consistent rolled up insights across portfolios - CEOs, CFOs and operating leaders using AI to analyze business performance - Sponsors and operating partners who see AI-generated numbers in board decks, lender updates and value creation plans - RevOps, sales and marketing leaders building Claude skills, custom GPTs or AI workflows on company data - Anyone who has shared an AI-generated number and later wondered whether it was right You'll leave with: - A clear way to explain to your team and board why AI analysis can be confidently wrong - Practical checks to tell whether your AI is working from complete, consistently defined data - A simple way to think about where AI belongs in your analytics stack, and where it doesn't About the speakers: Ed Marsh helps private equity sponsors and industrial and manufacturing companies drive organic revenue growth. He's an independent board director, NACD Directorship® Certified, and host of the Industrial Growth Institute podcast. Nikos Ntirlis is Partnerships Manager at Databox, an analytics platform that connects data from 130+ business tools into governed, consistently defined metrics that people and AI can trust. This is a conversation, not a sales pitch. Bring your questions. Tuesday, October 27, 2026 | 12:00 to 12:45 PM ET Streaming live on LinkedIn, YouTube and X. Can't make it live? Register at linkedin.com/events/confiden…

Your plant runs on OEE. Your revenue engine deserves the same rigor. Sales is essentially a factory. #OrganicGrowth #PrivateEquity

Marketing or sales: which matters more going forward? My answer is sales. In industrial markets, how you sell is the differentiator, not what you sell. Brian C. Reed of Anaconda says “it depends.” Here’s the debate. #B2BSales #Manufacturing