Craig Duxbury wrote a piece in The Drum that put a name on something the industry has been working around for years. He is a global B2B marketing and growth consultant, with 27 years of agency life behind him, and his argument is simple.

AI made execution cheap, and that exposed the absence of any real standard for client partnership. The premium has shifted to experience-based judgement, which was never properly defined or priced.

Read the piece if you haven’t. He’s right, and his core point lands. As soon as production gets fast and cheap, every weakness upstream becomes visible. His list: weak briefing, fuzzy accountability, soft commercial handling, poor strategic sell-in, creative dilution.

AI made execution cheap. That exposed the absence of any real standard for partnership.
A pair of worn leather boots — the boots of the CMO, who sees the same list from the other side.

Now put yourself in the shoes boots of a Chief Marketing Officer (CMO). He sees the same list from his side. The senior part of the work, the part that used to live in instinct and team knowledge, is now the only part that matters. In very few companies has this knowledge ever been made portable.

So far, I agree with Duxbury. Here is where I take the argument one step further.

Quality standards trapped in experts’ heads

A company standard for quality is a diagnosis in a document on your computer. It is a file that defines what good looks like as you see it. It is an artifact of consensus from your agency.

But a quality standard on a page does not live where the work is done. Senior judgement becomes valuable only when it stops being trapped in heads, calls, and group knowledge, and starts running where the work actually happens. That means it has to be embedded in the tools the team already uses: PowerPoint, email, AI tools. When judgement is portable, it is available after hours, in another timezone, the moment a regional team files a draft. That is portable judgement.

Embedded judgement vs non-embedded judgement

Consider a brand-voice review on first-draft content. A junior team member runs a draft through ChatGPT. The output comes back fluent and wrong. It sounds confident, but it is off-brand, the positioning is skewed, and the text has none of the regional nuance the team learned to add by hand three campaigns ago. The senior reviewer fixes the same five things they always fix. Again. And again. The judgement is human. It lives in one head, available during office hours.

The output comes back fluent and wrong.

Now picture the same draft run through a brand-voice review that is an AI skill, authored by that same senior practitioner. It is calibrated to the actual brand. It is signed by the person whose judgement it carries. The output is the kind of feedback that reviewer would give in person, in writing, in any region, in minutes. The senior person stops being the bottleneck for quality. The team stops sending fluent-and-wrong work to be checked. The brand stops paying for the inconsistency.

Three things have to be true at once for that to work. Otherwise it is just another tool. Embedded judgement has to be:

  1. Structured. A workflow with defined inputs, measurable outputs, and a quality bar you can audit. Not a prompt. A multi-step, high-quality skill.
  2. Authored. Signed by the practitioner whose experience it carries. Their name, with a version. Reviewed often, because the world keeps moving fast, especially in anything AI-related.
  3. Native. Running inside Microsoft Copilot, Claude, ChatGPT, Gemini. Inside the approved AI environment. No new login. No new platform. No migration project. And, most of all, not one more agent people have to remember and actively pull into the workflow.
When all three are true, knowledge is portable.

This brings us back to Duxbury’s commercial point. He says agencies need to stop giving away the highest-value parts of client partnership under the heading of “service.” I agree, again. And I would add one thing.

The service conversation only ends when the focus finally moves off pricing and onto the operational view. You have to put senior judgement where people actually do the B2B marketing. That is what drives improvement on the job.

So here is the question I would take back to your team this week. Which pieces of judgement in your group are currently team knowledge? Which senior reviewer fixes the same things in every project? Whatever the answer is, that is a good place to start the next improvement cycle.