The REEDS program · skill 3 of 5
Evaluate. Know what good looks like, and say so.
AI can produce a first draft of almost anything in seconds. Someone still has to decide whether it is good. Evaluate is the skill of knowing what good looks like, and saying so.
Book an intro call
What Evaluate means.
Evaluate is about judgment. It covers how you check work that AI helped produce, how you give feedback on it, and how you make decisions when the answer is not obvious and the stakes are real.
It draws on the group of capabilities MIT Sloan researchers describe as opinion, judgment, and ethics.
Polished output is easy to get and easy to trust too quickly. The people who add the most value are the ones who can tell strong work from work that only looks strong, and who will say which is which.
Why Evaluate matters more now.
When producing work gets cheaper, judging it becomes the job. Here is what the research says.
Polished output can still be wrong.
In an experiment with 758 consultants, AI improved speed and quality on tasks that suited it. On a task chosen to sit outside what the model of the time did well, consultants using AI were 19 percentage points less likely to produce a correct solution.
Source: Navigating the Jagged Technological Frontier (opens in a new tab). Dell’Acqua and colleagues, Harvard Business School working paper, 2023.
Confidence in the tool goes with less checking.
In a survey of 319 knowledge workers, higher confidence in generative AI was associated with less critical thinking. Higher confidence in their own skill was associated with more.
Source: The Impact of Generative AI on Critical Thinking (opens in a new tab). Lee and colleagues, Microsoft Research and Carnegie Mellon University, 2025.
Work that looks good and lacks substance is already common.
In a 2025 survey of 1,150 full time US desk workers, 40% said they had received "workslop" in the last month: AI generated content that looks good but lacks substance. They estimated about two hours to deal with each instance.
Source: AI-Generated "Workslop" Is Destroying Productivity (opens in a new tab). BetterUp Labs and Stanford Social Media Lab, in Harvard Business Review, 2025.
What Evaluate looks like at work.
Evaluate is a habit of asking "is this actually good?" before passing work along.
A confident answer that is wrong
An analyst gets a clean summary from an AI tool. They check two of the figures against the source before it goes to the client.
Feedback on work AI helped write
A manager gets a long, polished document. They give feedback on the thinking, not the formatting, and ask what the author would cut.
A decision with no clean answer
A leadership team has to decide where AI should and should not be used with customers. They talk about values before they talk about cost.
Evaluate at every level.
People learn a different side of Evaluate as they take on more responsibility. One topic per session.
Core · lead yourself
Judging AI assisted work
People learn to check AI assisted work before they rely on it: what to verify, what to question, and when to start again.
The Core level →Manager · lead a team
Feedback on AI assisted work
Managers learn to give useful feedback when AI did part of the work, and to hold a clear standard for what good looks like.
The Manager level →Executive · lead the organization
Decision making and ethics
Senior leaders work on judgment at scale: how decisions get made, who is accountable, and where the ethical lines are.
The Executive level →
Curious what REEDS could look like for your team? Let's talk.
Book an intro call
A 30 minute call with Dave Collinsof Oak & Reeds Interactive Training
Not ready to book yet?
Leave a note and Dave will reply to you himself.
