The Manager Skill for 2027 Is AI Judgment, Not AI Tricks

The manager skill that matters most with AI is judgment, not technique. Five calls cover it: delegate the right tasks, verify the output, disclose use to the people affected, protect confidential data and own the decision. Write the five calls down for your team, log decisions that involved AI and revisit the rules every quarter as the tools change.

I am Mac Malbrough. I hold a J.D. and an MS-HRM, I am not an attorney, and I help small employers prepare managers for new tools. This is a framework for judgment, so it will stay useful when the interfaces change.

What problem do managers face with AI?

Managers are caught in the middle. Leadership wants productivity gains. Employees want to know whether their jobs are safe and whether they are allowed to use the tools. The manager is expected to have answers and usually has a few weeks of experimenting and no policy. The instinct is to learn tricks: better prompts, new tools, shortcuts. Tricks go out of date quickly, and they do not tell you what to do when the tool gives a confident wrong answer.

What lasts is judgment about when to use the tool and when not to. A manager with good judgment and basic tool skill will do better than one with advanced tool skill and no judgment. That is the same pattern we see with any powerful tool, and it is why I teach the five calls first.

What is the first call: delegate?

Decide which tasks to hand to AI. A good candidate has a clear standard, a low cost of error and a person who can check the output quickly. Drafting a first version of a meeting summary fits. Writing a performance rating does not, because it is a decision about a person and requires evidence only you have.

The HR Task AI Readiness Scorer scores 30 HR tasks on four factors, so you can decide which to assist, which to pilot and which to keep human. A manager can adapt the same factors to the tasks on their own team.

What is the second call: verify?

Every AI output is a draft. Verification means checking the claims against a source, not rereading the text for fluency. A confident paragraph can contain an invented citation, a wrong number or a policy that does not exist. Build the habit of asking three questions before you use anything: what is the source for each fact, what did the tool leave out and what would I say if my name were on this?

For a task that matters, check against the primary document: the handbook, the statute, the contract. If you cannot verify a claim, remove it. Writing this down for the team removes the ambiguity: “you are responsible for what you send, whoever or whatever drafted it.”

What is the third call: disclose?

Decide when the people affected should know that AI was involved. A simple rule works for most situations: if the output affects a person’s pay, job, performance record or personal information, tell them and be ready to explain. For internal drafts, a short note that a tool helped is usually enough. This site follows the same practice. Website content is AI assisted.

Disclosure also covers the team conversation. Employees worry about AI replacing them, and silence makes it worse. My article on what to say when employees are afraid AI will replace them gives language for honest answers, including what you do not yet know.

What is the fourth call: protect?

Decide what data stays out of AI tools unless your company has approved them. A safe default is no names, no health or immigration information, no complaint details and no confidential financials. Check each tool’s data settings, including whether inputs may be used for training. Put the rule in writing and make sure every manager knows it. The California employee handbook template is where the policy belongs, and my safe AI prompt library shows prompts that work without personal data.

What is the fifth call: own?

A person is accountable for every decision, whatever helped make it. “The tool recommended it” is not a defense, and it is not a management style either. The AI Decision Log for HR records what the tool did, what a person checked and who decided. A manager who keeps a line for each AI-assisted decision builds a habit of ownership and a record in the same step.

For any decision involving hiring, discipline or pay, review the rules in your state. California has civil rights rules for automated decision tools used in employment, and I summarize them in the civil rights rules for AI hiring tools.

How do you put the five calls to work with the team?

Run a 30-minute team meeting with three parts. First, show the five calls on one page. Second, ask each person to name one task they would like to hand to AI and one they would not. Third, agree on how the team will check and disclose work. Write the outcomes down and revisit them quarterly.

Pair the meeting with development. Junior staff learn by doing the tasks that AI now handles, so managers must replace those learning paths on purpose, and how to train junior staff when AI does entry-level work covers that. For managers who are new to the role, the First-Time Manager Workshop Series offers a six-session facilitator kit, and the performance review and PIP templates keep review conversations grounded in evidence.

For the larger picture of using AI in HR, my book The HR AI Field Guide is available in paperback, with a Kindle edition.

Revisit the five calls every quarter. Tools add features, vendors change their data terms and your team will find uses you did not predict. A short quarterly review, with the decision log in front of you, keeps the rules aligned with what people actually do, and it gives managers a chance to share what worked.

What should you do this week?

Write the five calls as five sentences for your team, one per call, and bring them to your next team meeting. Start a decision log with your next AI-assisted task. After a month, look at the log and ask which calls came up most.

The AI Decision Log for HR is the fastest place to start. If you want help writing manager guidance for your company, book a free consult through my HR services for small businesses.

Related reading: where AI helps HR and where a person must decide.

Sources

• Small California employers and AI hiring tools: civil rights rules, with its regulatory citations
• Where AI helps HR and where a person must decide
• What to say when employees are afraid AI will replace them

McKinley holds a J.D. but is not a licensed attorney. Articles here are general information, not legal advice. For your specific situation, talk with an employment attorney.