Director of AI Innovation & Performance, Melbourne
I own performance strategy for an agency's largest accounts, and the data function the rest of the business runs on.
One half is the agency's largest accounts. Strategy, quarterly reviews, pitch work and the media investment decisions behind them, held against revenue rather than platform metrics.
The other half is the data function. Most agencies bolt AI onto the edge of what they already do and call it transformation. I am building the layer underneath: consistent measurement, one definition of performance across paid and organic, and reporting senior people will actually make decisions on. Automation only compounds on top of that.
Every build carries a business case. Build cost, run cost, maintenance, and the billable hours it returns. A good share of what gets proposed does not clear that bar, and killing those is worth as much as shipping the rest. What does get built is built with the team who will own it, because a tool nobody adopts is a write-off.
Each of these carried a business case before a line of it existed, and each is measured on the hours it returns to fee-earning work.
A proper account review costs a specialist the better part of a day: hundreds of settings, thousands of keywords, and very little of it judgement. This does the reading in minutes, checks the 243 things a strong specialist would check across Google, Meta and Microsoft, and writes both the findings and the client deck. What it does not do is decide. That stays with the strategist, which is the only part of the job worth paying for.
Ad platforms, analytics, Search Console and Merchant Center sit in separate systems, so nobody sees the whole picture. This answers one question across all of them in plain English. Read-only by design, so it can never touch a live account.
Leadership had four reports and no answer. Every client now carries one score built from advertising, SEO, budget position and satisfaction, so underperformance surfaces before the client raises it.
Overspend used to surface on the invoice, and the credit note followed. Every account is now checked against budget each morning, and drift is flagged while there is still margin to correct it.
Every account quietly funds searches nobody intended to buy. This reads the search terms daily, isolates the waste, and returns a list the team approves in one click, with the saving tracked rather than estimated.
Onboarding a client used to cost an hour of an account manager's day. One click in Gmail now builds the project board, drafts the replies, and turns meeting notes into assigned work.
Client reporting was assembled slide by slide, every month, for every account. The numbers, charts and commentary now land in a finished deck, so the time goes into the recommendation instead.
Smaller pieces, all in service. A content planner that reads a client's site and the live search results, then drafts the brief, the questions and the page copy for around twelve cents a page. Measurement for the traffic now arriving from AI assistants rather than search engines. A handover tool that reads a client's full email history and reports where the account actually stands. And a Mac app that folds my calendar, tasks and meeting notes into one view.
Nine years, from running campaigns to running the systems a whole agency depends on.
The account where I hold every lever at once, and the only honest test of anything I would put in front of a client.
An Australian pet supplies business I started and run on my own. Store, catalogue, product feed, paid media, SEO, email and the margin behind all of it. Running a P&L rather than a retainer changes how you read a number. It is where a tactic gets proven, or quietly killed, before it ever reaches a client's budget.
Visit fluffrepublic.com.auSixteen sectors across ecommerce, lead generation and considered purchase. The mechanics change from one to the next. The commercial questions do not.
What I work with, on both sides of the role.
ADMA, August 2023
ADMA, December 2024
Australian Institute of Management, October 2022
RMIT University, Melbourne
Feb 2019 to Nov 2020
Software engineering, project management and usability engineering. The reason the building half of the job came naturally.
University of Mumbai
2015 to 2018
Economics, management accounting and trade finance. Where the commercial lens came from.
Strategic partnerships, advisory work, and how AI actually lands inside an agency rather than how it demos.
Connect on LinkedInArrows to move, Enter to go, Esc to close