Management, not tools

I handed over a brand system in November. Full file sources, component library, tone guide, photography brief, signage templates, the works. The person who commissioned it left three weeks later. The new hire inherited the login, opened the folder, and froze. Not because the system was complicated. Because no one had taught them what decisions they were allowed to make, what needed sign-off, or when to use which asset.

The technology was fine. The management layer was missing.

Why does management matter more than the tools?

Because the tools do what they are told. A brand system will not enforce itself. A content management system will not write the copy. A campaign template will not know whether the site hoarding goes up Friday or the pitch document is due before the planning submission. Management is the layer that decides what gets done, who does it, and whether it is good enough to ship. Without that, the best-documented handover in the world just sits in a folder.

Ethan Mollick, who researches AI and work at Wharton, recently argued that management is becoming an AI superpower. His thesis: agents and models amplify whatever capability you already have. If you know how to structure a task, delegate clearly, and evaluate output, AI makes you faster. If you don’t, it makes you slower, because now you are managing both the work and the tool that is supposed to help.

I see this every time I train a project team. The ones who already know what a good site photograph looks like can art-direct an image model in two tries. The ones who don’t will generate forty variations and still pick the wrong one. Same model, different outcome. The skill is not prompting. It is taste, and the ability to articulate it.

What does a brand system look like for a specialist business?

A surveying practice needs to update the capability brochure for a new service line. They have the brand system, the photography library, the layout templates. Do they know which image to lead with? Whether the tone should shift for this client type? When to ask for sign-off versus just shipping it? If the answer is no, the system will not help them. If the answer is yes, they will turn it around in an afternoon and it will be on-brief.

Or: a developer is preparing a pitch document for a local authority brief. The director is in three meetings. The associate has access to everything: financials, site photography, previous submissions, a working content management system. Can they shape a narrative, write a cover note, and export a PDF that does not need a full rewrite? That is a management question, not a tools question. The associate either knows what the director cares about, or they don’t.

This is why I structure handovers the way I do. Documentation alone does not transfer capability. You also need the editorial walkthroughs, the on-site training, the worked examples. Not because the system is complicated, but because someone has to learn how to manage it. What gets updated when. Who makes the call. What good enough looks like on a Tuesday afternoon when the contractor capability statement is overdue.

What happens when the management layer is missing?

The system decays. Not immediately, but predictably. The photography style drifts because no one remembers the lighting brief. The tone flattens because no one is holding the line on clarity. The content management system fills up with drafts no one signed off on. Six months later, the brand looks like it was designed by three different agencies, and the team calls me back to refresh it. The system was fine. The discipline was not there.

McKinsey research on digital transformation found that technology adoption rates are high, but utilisation rates are low. Translation: people buy the tools, then don’t use them properly. The gap is not technical. It is organisational. Someone has to own the process, set the standard, and make sure the work actually ships.

This pattern is well documented in design systems literature. According to Nielsen Norman Group research on design system adoption, the primary failure mode is not technical debt or incomplete documentation. It is the absence of governance: no one owns decisions about when to extend the system, when to follow it strictly, and when exceptions are allowed. The tools are necessary but not sufficient.

Harvard Business Review research on organisational capability reinforces this: systems and processes mean nothing without the discipline to use them consistently. Culture eats strategy for breakfast, and management eats tools for lunch.

This is true for brand systems, and it is about to become more obvious with AI. A model can draft pitch copy in four seconds. But can the associate tell if it is on-strategy? Can the project manager brief it clearly enough that the first draft is usable? If not, the speed advantage disappears. You will spend more time fixing bad output than you would have spent writing it yourself.

How should a small specialist business adopt AI in 2026?

Start with management capability, not the technology. Work out who on the team can already evaluate good work, articulate a clear brief, and make a call without three rounds of sign-off. Train them on one model, one task, one workflow. Let them get fast at that before adding more tools. The goal is not to adopt AI everywhere. It is to adopt it where management capacity already exists, so it actually improves the output instead of just creating more drafts to review.

I run AI automation sessions for project teams, and the structure is always the same. First hour: no tools. Just working out what decisions you actually make, who makes them, and what good looks like. Second hour: one model, one task, live. The teams that already have clarity on the first part move quickly. The ones that don’t will struggle regardless of which model they use.

This is why I frame it as: AI arms teams, it does not replace them. The armed team still needs to aim. If they can’t, the weapon does not help. You can read more about how I approach brand work for specialist businesses and the handover structures that make systems usable after the studio leaves.

What should a business document before the agency leaves?

Not just the file sources and the fonts. Those matter, but they are not enough. Document the decisions. What gets signed off by whom. When to use which template. What tone applies to which audience. How to brief external suppliers. What the photography standard is, and how to evaluate whether a new shot meets it. When to update the content management system versus when to leave it alone. What gets published immediately versus what waits for the quarterly refresh.

The teams that can run a brand system after I leave are the ones who inherited that layer, not just the assets. The assets are easy to hand over. The decision-making structure is harder, and it is the part that determines whether the system actually gets used.

I cover this in detail in my post on what to document in a brand handover, and you can see how it plays out in practice in recent project case studies.

Does this mean the tools don’t matter?

No. The tools matter. A well-built content management system is faster and clearer than a folder of InDesign files. A component library is easier to maintain than a set of one-off templates. A structured photography brief produces better output than “get some nice shots”. I spend a lot of time choosing and building the right tools, because they set the ceiling for what the team can do.

But the floor is set by management. A great system run badly will underperform a simple system run well. I would rather hand over a tight process with basic tools than a complex platform with no one who knows how to manage it. Capacity is the lever, not the infrastructure.

If you are preparing for a handover, planning a rebrand, or working out how your team should start using AI this year, send a paragraph. I will write back, usually same day.