UpTrajectory Review

Dessy Pavlova's piece in The Next Web tackles a persistent and costly myth: that AI can repair a business that has never bothered to streamline its processes. The available excerpt is brief, but the argument is clear enough — AI can accelerate operations, automate repetitive work, and surface problems quickly, but when you feed it into a fragmented or poorly designed workflow, you are not solving anything. You are scaling the dysfunction. That framing alone makes this worth reading, because it cuts against the prevailing narrative that AI adoption is a cure-all for operational chaos.

For a small-business operator, this is not an abstract concern. Most small businesses run on processes that were never formally designed — they evolved out of habit, staffing constraints, and whatever worked in the moment. The owner who thinks an AI tool will finally organize the quoting process, the inventory handoffs, or the customer onboarding sequence is likely to be disappointed. AI does not impose order on disorder. It reflects whatever structure, or lack of structure, already exists. If your intake process is inconsistent, an AI tool will automate that inconsistency at scale.

What is genuinely useful here is the refusal to let AI vendors off the hook. The piece implicitly challenges the assumption that the technology itself is the bottleneck. In practice, the bottleneck is almost always upstream: unclear decision rights, redundant steps, undocumented tribal knowledge, and workflows that exist because nobody ever questioned them. This is not a new insight in process management circles, but it is under-reported in mainstream AI coverage, which tends to focus on capability rather than fit. We agree with Pavlova's skepticism. The businesses that benefit most from AI are rarely the ones that needed it most.

The second-order effect worth considering is cost. Fragmented processes do not just waste time — they compound. When AI is layered onto a broken workflow, the business often spends money on tools, training, and integration before discovering that the output is unreliable or that the automation is producing errors faster than humans can catch them. There is also a cultural cost: employees who already distrust the process become more resistant when a new tool appears to be making things worse. The downstream effect is that AI adoption can actually delay the harder, more valuable work of process redesign.

What to watch next is whether this argument gains traction beyond opinion pieces. If more businesses start auditing their workflows before selecting AI tools, vendors will be forced to change how they sell — from promising transformation to promising fit. For now, the practical takeaway for operators is straightforward: before evaluating any AI solution, map the process it is supposed to improve. If you cannot draw it clearly, no tool will fix it for you.

Pavlova's piece is a useful corrective at a moment when AI spending is accelerating faster than process discipline. The excerpt is only a preview, and the full argument likely goes deeper into specific failure modes. But even the available text is enough to prompt a hard question that most small-business owners would rather avoid: is the problem you want AI to solve actually a technology problem, or is it a process problem you have been deferring?

“AI can accelerate an operation, automate repetitive work, and surface problems at remarkable speed.” — The Next Web

Takeaway: Before buying any AI tool, map the process it is meant to fix — if you cannot draw it clearly, the tool will only scale the confusion.