RLK Unfiltered
Not a list of things you could Google on your own. Perspectives written with the same rigor applied in client engagements, specific enough to be useful, direct enough to share with your team.
Producing work got nearly free while checking it still runs at human speed, through people whose judgment cannot be bought in bulk. That is where AI programs back up, at every company size. Here is the math on the jam and what clears it.
The SOC 2 arrived on time, the box got checked, and none of it covers the model. Why AI just moved the burden of proof to your side of the table, and the contract terms that buy back visibility.
So how do you reconcile being a technology consultant when AI data centers are drinking all our water? A real answer with real numbers, the five questions your vendors should sweat, and the contract terms that turn water pledges into consequences.
Is anyone actually making money with AI? Somebody is, and it's probably your vendor. Where the gains really sit, why the savings won't stay, and what builds revenue that lasts.
AI rewards speed and few approval layers, which is exactly what small and mid-sized companies have and the Fortune 500 does not. They can do more with AI today than the giants. The problem is access: the firms that could help are priced for the enterprise, so the companies best built to sprint with AI are the ones locked out of the help.
Most AI board presentations fail because they present AI as a technology update. Boards do not fund technology updates. They fund financial outcomes. The slide that works is the one that shows a dollar number, a baseline, and a timeline.
The agent wins on sticker price. But the salary was never buying year-one output. It was buying your 2031 senior team, and the apprenticeship deal that produced it just collapsed. Here is the math, and the opening it creates.
Machine identities outnumber your humans 82 to 1, and AI agents are turning those credentials into coworkers with no manager and no offboarding. An airline already tried blaming one in court. It went badly.
Companies that buy AI from vendors succeed far more often than the ones that build it in-house. Yet mid-market teams keep spending engineering on capabilities a vendor already sells, then buy their one real differentiator off the shelf. Here is how to tell the difference.
AI buying decisions are being made by the people who understand the technology. They should be made by the people who understand the money. Your CFO is the most undertapped AI leader in your company, and the reason is structural.
The pilot worked. The demo was impressive. Leadership is excited. And in six months, nothing will have changed. The gap between a successful AI pilot and captured economic value is where most companies lose, and almost nobody is talking about why.
The companies that say they can't figure out AI can usually figure out AI just fine. What they can't figure out is who is allowed to say yes. Authority structure, not tool selection, is the binding constraint on AI value capture in the mid-market.
The phrase is doing damage. When a program is called AI strategy, the team optimizes the AI. The strategy lives at the workflow. The AI is the instrument. Programs that move numbers are organized around the workflow. Programs that produce activity reports are organized around the technology.
The simplest way to make sense of Sam Altman: he is a large language model from ~2034, fine-tuned on the complete works of Y Combinator, and instructed to build the very thing he keeps warning us about. The evidence is in the outputs.
The MIT finding on AI pilot failure is a verdict on procurement. The 5 percent that worked share a structural posture toward scope and accountability, and that posture is teachable.
The Gates Foundation's malaria intervention points to a deeper lesson: philanthropic capital creates far greater impact when deployed with analytical discipline: the same rigor that governs effective investment capital.
People across organizations use AI constantly, then remove every visible trace of it. The story is a professional norm forming in real time, and it carries a real cost.
You can have a sharp strategy and still fail to execute it. In most organizations, the gap is measurement, and the board dashboard is where it shows.
An operating model is how work flows from idea to production and who is accountable at each step. The diagram is decoration. Here is what actually matters.
Valuation rewards outcomes, not activity. Here is what the data actually says about value creation and where companies routinely overspend with little return.
C-suite transitions are deceptively fragile. Up to 40% of senior executives fail within 18 months. The difference is sequencing.