The management skills AI makes mandatory, not optional

When the planning committee is three weeks out and the AI has drafted six versions of the community benefit statement, someone still has to choose the right one. When the leasing brochure needs a refresh and the system generates four layout options in twenty minutes, someone has to know which one works for the building and why. The management skills for AI are not new. They are delegation, quality control, taste, and the ability to articulate a standard before you ask for the output. I see this in every training and handover engagement I run. The teams that adopt AI well are not the ones with the best prompts. They are the ones who already knew how to brief, review, and make a call.

What are the core management skills for AI work?

Delegation, quality assessment, and clarity of instruction. AI cannot infer what good looks like from a vague brief. It needs a standard, examples, constraints, and a sense of the audience. The management skills for AI mirror the ones required when working with any capable but literal collaborator: clear scope, quick feedback, and the confidence to reject work that misses the mark. Ethan Mollick, a researcher at Wharton, describes this in his One Useful Thing essay on management as superpower, noting that managers who can delegate, evaluate, and iterate outperform those who try to do everything themselves, AI or not.

Why does quality control become more important, not less?

Because volume goes up and the cost of bad output drops to near zero. When a system can generate twelve versions of the investor update in an hour, someone has to catch the one that uses the wrong property examples, overstates the yield, or misreads the tone for the investor audience. Speed makes quality control the bottleneck. A 2023 BCG study on generative AI and professional work found that consultants using generative AI completed 12.2% more tasks but quality dropped when oversight was weak. The skill is not running the tool. It is knowing what to reject before it goes in the board pack.

How does delegation change when the first draft is instant?

The brief becomes the work. When you can generate a pitch document in fifteen minutes, the time saved moves upstream into defining what the document should argue, which lenders or investors to address, and what the property story actually is. Delegation to AI requires the same clarity you would give a junior team member on their first week: context, constraints, examples of good and bad, and a clear definition of done. I see this in every system handover. The teams that struggle are the ones who never had to write a brief down before.

What does knowing what good looks like actually mean?

It means you can articulate why one leasing brochure works and another does not, in thirty seconds, without sending a redline. It means you recognise when the tone is wrong for the tenant mix, when the property examples do not match the building type, or when the layout will not work at A3 print. Taste is a technical skill. It is pattern recognition built from repetition, mistakes, and seeing enough examples to know what separates competent from correct. AI makes this skill load-bearing because it removes the production bottleneck. You are no longer judging the best output a human could do in two days. You are judging twelve options produced in twenty minutes.

How do you train a property team to work this way?

By making the handover the project, not an afterthought. Every brand system, CMS, or content operation I build is documented for the next associate to run without me. That means file sources, tone guides, example briefs, content rules, and recorded walkthroughs of the system in use. I train on real scenarios: the next hoarding brief, the investor update due Friday, the leasing brochure refresh that has been on the list for two months. I do not teach prompt libraries. I teach how to evaluate output, how to brief with specificity, and how to reject work that does not meet the standard. The goal is not AI literacy. It is decision-making speed.

What happens to teams that do not build these skills?

They produce more, faster, and worse. The system generates plausible mediocrity at scale and no one catches it until the investor calls, the planning committee pushes back, or the leasing agent says the brochure does not match the building. Volume without judgment is not productivity. It is risk. A 2024 Stanford HAI study on AI adoption in professional services found that performance gains were highest among workers who combined AI use with strong domain expertise and critical evaluation skills. The ones who outsourced judgment outsourced quality.

Where does this show up in a property context?

In every situation where speed matters but correctness matters more. The funding proposal that needs to land with three different lender types. The planning submission that has one shot at committee. The partnership brand that has to work for both sides of the capital structure. The investor report that goes to a family office principal who reads every word. AI gets you to the first draft faster. The management skills for AI get you to the version you can actually send.

How do I build this into studio practice?

By treating capability transfer as a deliverable, not a courtesy. Every engagement includes documentation, training, and a handover session where the internal team runs the system with me in the room. I do not hand over a Figma file and a login. I hand over the decision framework: what to brief, how to evaluate, when to reject, and how to maintain the standard when I am gone. AI is in the production layer of almost everything I do, but I do not sell AI advisory or run workshops on prompt craft. I train teams to delegate, evaluate, and ship with confidence. The tool changes. The skill set does not.

If your team is starting to use AI for marketing, content, or investor communications and you want to build the management layer properly, send me a paragraph on the situation. I will write back with what the handover might look like.

Frequently asked questions

What management skills does AI require?

Delegation, quality control, and the ability to articulate a standard before asking for output. AI works best when treated like a capable but literal collaborator who needs clear briefs, examples, and constraints.

Why is quality control more important with AI?

Because AI can generate volume at near-zero cost, making oversight the bottleneck. Without strong quality control, teams produce more work faster but at lower quality, which creates risk rather than productivity.

How does AI change delegation?

It makes the brief more important than the execution. When a first draft takes minutes instead of days, the time saved moves into defining what good looks like, what the audience needs, and what the output should achieve.

What does taste mean in an AI context?

The ability to recognise why one option works and another does not, quickly and without detailed explanation. Taste is pattern recognition built from experience, and it becomes essential when AI removes the production constraint.