Every week, dozens of emails, LinkedIn posts, and articles land in our inboxes — the latest model benchmarks, new AI tools, another startup acquisition. Fair enough; I read them too, it's part of the job. But how many of those updates actually matter to business decision-makers?
The reality I see on the ground every day looks quite different. The questions managers are actually asking are things like:
Where do I start when I have dozens of manual processes and no idea which ones to prioritize?
Should I train all my staff on AI, or focus on a few dedicated people?
Are my internal data good enough to feed an AI tool — and how do I even check?
If the AI makes a mistake, who's liable — the tool, the vendor, or my team?
How do I measure ROI on an AI project before rolling it out company-wide?
To tackle these questions, you need to step back from the noise and return to a first principle: AI is a tool — a powerful one when used appropriately, but a tool nonetheless. Your business strategy doesn't change because of AI. It gets sharper when AI is properly integrated into it. The right questions remain: What are my company's strengths and weaknesses? How can AI reduce friction where I'm vulnerable, and amplify the advantages I already have over competitors?
That's the lens this newsletter is built around.
A CEO told me this recently. His plan: three or four tech-savvy employees, trained thoroughly, managing all AI initiatives. Clean and contained.
He was half right. Not everyone needs deep AI skills — a small group owning technical implementation makes sense. But assuming the rest of the company can stay AI-illiterate is where it breaks down.
When only a few people understand AI, adoption fails quietly. Others don't trust outputs they can't evaluate, work around the tools, or use AI results uncritically. The "AI team" becomes a bottleneck.
Basic literacy needs to be company-wide: what AI can and can't do, how to prompt effectively, how to spot a wrong output. A half-day workshop, not a six-month curriculum.
And the CEO himself? Non-negotiable. A leader who doesn't understand AI can't prioritize projects or assess ROI claims — and becomes dependent on whoever holds the technical knowledge. That's a governance risk.
When spreadsheets arrived, not everyone became a financial modeler. But everyone learned to read one. The CFO especially.
He's now planning a two-track training program.
Arya International Trade (Ningbo) manages thousands of spare parts SKUs across dozens of vendors on ZOHO. Every vendor price update meant manually cross-checking an Excel file — thousands of lines — against ZOHO inventory. The complication: vendor part names never matched ZOHO references exactly, and the same part could appear from multiple vendors. One update: up to three weeks of work.
The fix: a two-step automated workflow. Upload the vendor file via a simple web interface, enter the vendor name and exchange rate, receive a color-coded Excel comparison ready for decision-making. One final upload pushes approved prices directly into ZOHO.
Three weeks → one hour per vendor
Full ROI in one week
Staff relocated from data entry to sales — with direct revenue impact
"The automation from AiGAIN is amazing. We reduced our update time from weeks to only one hour." — Yannick Le Hellaye, CEO, Ningbo Arya International Trade
Don't ask "where can we use AI?" Ask "where does it hurt?" List your five most time-consuming or error-prone manual processes and run each through three questions:
Is it repetitive and rule-based?
Do we have the data to feed it?
What is it actually costing us?
The process that scores highest on all three is your pilot.
Example: An accounting firm automated invoice processing — 40 PDFs a day re-entered manually — instead of the flashier client reporting project. Two weeks to deploy, one month to ROI. The time recovered went straight into higher-margin advisory work.
In the spring of 2022, the Oklahoma City Thunder sat near the bottom of the NBA standings. Three seasons later, they were champions… →
Most discussions about AI and management focus on the wrong question. The issue isn’t whether AI will eventually replace managers. It’s what happens when managers stop acting like managers before that question is even settled. →
The debate about AI and employment has been asking the wrong question. Organizations ask which jobs will disappear. Employees ask whether they’re next. What neither side is asking — and what actually determines outcomes — is this: which tasks are changing, and what does that free people to do instead? →
Most organizations are fighting the wrong battle. They’re spending the majority of their AI budgets on vendor evaluations, model comparisons, and pilots — and wondering why ROI remains elusive. →