How the AI content deluge will make decisions harder

AI is about to reshape the work of those who make decisions, with a big impact and the rest of the workforce that acts on those decisions. Why? Because it will generate a flood of suspect “insight” that will overwhelm decision-makers.

Reports and analysis fuel decisions

The fuel that drives all rational corporate decision-making is the report. Decision-makers analyze data, compile it in reports, share it with managers and executives, and act on it. Once they’ve made a decision, rank-and-file workers in departments like marketing, sales, service, IT, and product development act on the content in the report.

It has been this way for a long time.

Reports have always been time-consuming to create. Technologies like the internet, word-processors, data warehouses, and business intelligence tools have made it easier to create these reports, but the tools cannot substitute for the insights of experienced and intelligent analysts. If a report was to be useful, it needed an analytical thinker behind it.

Organizations like Netflix and Amazon have embraced written reports (sometimes referred to as memos) as fuel for making decisions, prioritizing them over endless meetings. These organizations require anyone who would like to make a change to pull together a case, acquire and analyze data, consider alternatives, and demonstrate how the change will impact a key metric like sales, costs, or profits.

The speed of decision-making was always gated by the time-consuming work of analysis and synthesis, and the limited number of staff who could think and write strategically.

AI is about to change that forever.

Insights and reports in the age of AI

Now that AI can write large volumes of content easily, reports are no longer a bottleneck. Anyone can tell ChatGPT, Gemini, or Copilot “help me draft a report on this.” These AI tools will answer questions about where the corporate and public data is, how to interpret it, and how to format it into a report intended to fuel a decision.

But just as inexperienced engineers “vibe-coding” applications are creating inadequate, poorly engineered, insecure, and crappy applications, inexperienced analyst-wannabes can now create an arbitrary quantity of bullshit reports.

Such reports will look like the reports that came before. AI can do that easily. They’ll be structured like the old reports. They’ll be similarly focused on data. They’ll have similarly formatted recommendations.

Will the reports be any good? I worry about that.

These AI tools are designed to make those who create them happy. Delighting a report creator does not serve the goal of creating a wise, balanced, accurate, and strategic report.

These reports may be full of fabricated information and inappropriate sources. Inexperienced analysts cherry-pick data to support their conclusions. AI makes that far easier.

Inexperienced report-writers focus on volume. But longer reports are worse, not better. An inexperienced analyst doesn’t know what to include and what to leave out, so they’re likely to leave almost everything in.

At some point, the corporation needs a human needs to make a decision (I hope). Launching a product, changing a process, acquiring a company — these decisions and those like them have potentially massive consequences.

Those human decision-makesr are now going to be deluged with far more authoritative looking reports from their ambitious staffs. I expect massive decision fatigue. I expect errors. They’ll have pressure from above to act more quickly and decisively — after all, now they’re empowered with AI-fueled analysis, however dubious.

When it’s easier to write, the increase in volume will make it harder to read critically. This the read-write imbalance that’s about to hit every corporation. Laying people off to take advantage of “AI efficiencies” will only exacerbate the problem.

Inevitably, such decision-makers will fight back with their own AI tools to survive the flood — which will only amplify the AI-created biases and decision errors. Their analytical and critical-thinking skills will take a back seat to AI-powered tools that simply keep their heads above water.

Look for the signs

Already, we see headlines about poor decisions blamed on AI.

UnitedHealth was accused of using AI to deny care that was recommended by health professionals.

Duolingo backtracked on using AI in performance reviews.

But as far as I know, no one has yet admitted that an AI-created report led to a bad decision — even though, I believe, that’s likely already happened thousands of times throughout the corporate world.

If I’m right, the flood of bullshit “analysis” is going to demand a complete overhaul in corporate decision-making.

I don’t know the solution to this problem. But I know this much so far. The absence of a strategic perspective on this issue is going to make workers’ jobs harder, make decision-makers more fatigued, and make smarter decisions harder, not easier.

More reports is not equivalent to more wisdom. The sooner decision-makers find a way to harness AI properly, rather than just using it to create more report volume, the better off they’ll be.