I realise I've been away for what seems like forever. The truth is, life, work, pretty much everything has been hectic.
In the last couple of years, unless you've been on a remote island with no internet, the marketing industry has been undergoing a huge
shift towards technology-powered enablement. With new demands, work is faster-paced and more demanding than ever.
So when I recently stepped in to support a global client account facing a perfect storm — a massive surge in deliverables, complex
multi-market regulations, multi-language requirements, and zero time to scale a traditional team — I knew that experience alone wasn't
going to be enough.
The challenge wasn't just the creative. It was the variables.
With country-specific clearance rules, rate cards, formats, and language combinations, the estimating and approval process alone was
a minefield of delays and errors. I realised that years of client services training, as valuable as they are, weren't sufficient in
today's rapidly changing, technology-led landscape. It was essential to put a repeatable system in place — and fast. Not in the
conventional way you'd normally build out a team, because we didn't have the luxury of time. We needed something easy to follow, with
minimal onboarding, so that anyone joining the team could fast-track their way into the account from day one.
Here is the three-step system I built.
Step 1: Rapid onboarding — the knowledge build
The first challenge was getting up to speed without weeks of traditional onboarding. I turned to a secure enterprise AI platform
to build a custom agent for this account — uploading relevant account documentation, operational guidelines, and market-specific business
rules to create a knowledge base the agent could draw on immediately.
What would normally take weeks of briefings, file reviews, and relationship-building compressed into hours. The agent could reference account context, flag market-specific considerations, and surface relevant precedent — all within a secure environment that kept proprietary information contained.
The speed wasn't the only benefit. The consistency was. Every team member accessing the agent got the same foundational knowledge, not a version filtered through whoever happened to be available to brief them.
Step 2: Codify the process — create a repeatable script
Getting myself onboarded was only the first problem. The estimating process itself — cross-referencing media plans with country-specific rate cards, format variables, language requirements, and clearance rules — was where the real complexity lived, and where errors were most costly.
I didn't have the luxury of building a formal system from scratch. We were already in the middle of delivery. So I did something more immediate: I encoded the business rules into a detailed prompt.
Not a vague instruction. A repeatable script and specific enough that the agent would synthesise media plans with rate cards, formats, and language variables consistently, every time, regardless of who was running it. The prompt became the process. The process became the standard.
This is a discipline I'd encourage any senior practitioner to consider: the knowledge in your head about how a process should run, the rules you apply intuitively after years of experience — can that be written down precisely enough to become a prompt?
In most cases, yes. And doing so is one of the highest-value things you can do with AI right now.
Step 3: Scale the capability — the team force multiplier
Getting myself up to speed was half the battle. To hit deadlines, I gave the wider team access to the agent and the estimating script — so the expertise could travel instantly, without me needing to be in every conversation.
The team could now:
- Generate accurate estimates by cross-referencing media plans with complex, country-specific rate cards and language requirements
- Navigate advertising clearance and market-specific regulations with far fewer false starts and rework cycles
- Onboard onto the account in hours rather than weeks, with a consistent foundation of knowledge from day one
The shift I was making was deliberate: from being the person who does the work to being the person who oversees the system that does the work. That distinction matters enormously in high-pressure, high-volume environments.
---
The result
We didn't just survive the surge. A team that hadn't spent any time on the account could deliver content at scale and produce accurate, complex estimates in minutes — scoping markets correctly from the start, rather than discovering errors after the fact.
The system held. The deadlines were met. And the knowledge that had previously lived only in the heads of experienced team members was now accessible to everyone, consistently.
What this taught me about AI and lived experience
Around the same time, I came across the term **"lived experience"** in a strategy context I was working on. It stayed with me — because it captures something important about what I'd actually done.
My client services knowledge — the operational judgment about how to scope a multi-market campaign, how to read a rate card, how to
anticipate where a clearance process will create delays — has been earned through years of doing this work. It's not theoretical. It's the kind of knowledge that comes from having been in the room when things went wrong, and adjusting accordingly.
What AI allowed me to do was take that lived experience and make it distributable. Not perfectly — an agent doesn't replicate thirty years of professional judgment. But it captures enough of the methodical, rule-based layer of that knowledge to make a junior team member significantly more effective, immediately.
That, I think, is the real promise of AI for senior marketing professionals. Not efficiency. Not automation. Multiplication.
Two kinds of work
I've come to think about the impact of AI in agency environments through a simple distinction:
The methodical — the maths, the variables, the compliance rules, the data cross-referencing — this is work that AI can handle consistently and at scale, once it's properly encoded. It doesn't require judgment. It requires precision, and AI is better at precision than humans who are under pressure and short of sleep.
The intuitive — client relationships, strategic direction, creative judgment, the ability to read a room and know when the brief isn't quite right — this is where human experience creates the edge that no agent can replicate. This is where thirty years of practice actually lives.
The most valuable thing a senior marketing professional can do right now is be honest about which parts of their work fall into each category — and start encoding the methodical ones, so their actual thinking can focus on the intuitive ones.
---
The question worth sitting with
When you don't have time to build formal systems — and in agency life, you rarely do — you still need to find a solution fast.
Building AI agents lets you capture what your most experienced lead knows and distribute it across a team instantly. It's not a replacement for expertise. It's a way of making expertise less dependent on any one person being in the room.
In high-pressure agency environments, the goal isn't just to be faster. It's to build systems that allow your team to be as smart as your most experienced lead, from day one.
The question I'd leave you with is this: **what do you know, that your team doesn't yet have access to — and could an agent change that?**
