I am an operator, not a vendor, and I have no product to sell you. I spend three days inside your finance and operations team, find the work worth automating, and build it on the systems you already run.
People have built things ("skills"), plenty of them. Here is what I walk into almost everywhere.
Every one sits in somebody's personal account, where nobody else can reach it.
Each one was built a different way, so the same question asked twice gets two different answers.
Nothing gets properly tracked, and nothing handles its own errors.
The models are good enough. What is missing is one owner, one baseline, and one written rule for what the machine may do without asking.
How your team actually closes the month, held in one place instead of in three people's heads. I call that a second brain.
The longest-serving member of your team retires. The reason that entry is booked the way it is does not retire with them.
Everything I build for a client, I have run on myself first.
See the buildsThousands of card transactions a month, each read by a person before approval. The team was about to shop for a tool to help with the pile.
The expense platform had shipped a policy checker months earlier, on the tier they already pay for. It tests every transaction against the written policy.
No new tool. Turn on the feature, and build it into a tracked workflow with proper constraints. Coding judgment still stays with people.
The capability ships inside a product you subscribe to, switched off.
Two systems that could talk on licences you hold, and a person filling the gap every week.
Macros, a script, an AI assistant on one person's account. It breaks quietly, and the output still looks right.
Most companies have them. An agent answers a question and stops there.
Few have one: a named owner, a baseline, and a written rule for what the machine may touch.
Almost none have one: your rules and decisions in one place, read by every agent.
Grant Thornton, 2026 AI Impact Survey. 100 private equity fund leaders, February to March 2026.
Three days inside your team, then a written readout.
$18,500, fixed. The findings are yours, whether you continue with me or not.
One or two days a week, with the person who will run it after I go.
Priced per agent in production, not per hour of me.
| Instead of | You get |
|---|---|
| Three days of copy and paste | A monthly package pulled from the ERP and the bank feed, commentary drafted before the controller opens the file |
| The spring audit scramble | Evidence matched to the request list as transactions happen |
| The accountant everyone waits for | Reconciliations, accruals and intercompany matched in shadow mode until the machine agrees with your people |
| Asking the one person who knows | How the company runs, written where a machine and a new hire can both read it |
| A pilot described as promising, for a second year | A weekly ledger of hours returned against your baseline |
If your ERP already does one of these, I will tell you to switch it on, and I will go home.
Your data will not carry an agent yet. Building on it automates the mess.
A feature you pay for every month, switched off. Turn it on and keep the money.
Put a name on it first, or there is nothing for a build to move.
The agent runs, your owner runs it, and there is no reason to keep paying me to watch.
I have owned the process, run the team, and answered to a CFO for the result. The same pattern every time: understand before you buy, or build.
Attrition from 25% to 12% while raising price.
Built, then ran, the world's first certified pre-owned phone program and the first trade-in program run by a phone maker. $100M+ each.
Head of finance, planning, procurement and logistics: $150M+ budget, over 30,000 doors, 700 FTEs.
Built, then ran, a global trade-in business inside an insurer that had never had one. Three countries, two continents.
A product line losing money on every unit. Nothing was built: they widened what they traded, and revenue and gross profit both went up.
Five weeks, five deliverables. The biggest item on their automation list was a capability they already paid for, switched off.
CELX Advisory started in 2021 after Samsung, advising in reverse logistics and trade-in. In 2026 it became the home for AI implementation, because that is what operators were asking for.
CSC (now DXC Technology), TXU Energy, BCG, Samsung, SquareTrade. I have spent my career at the front edge of the tools, and I take the part that is useful and leave the rest. I can hold the argument in the boardroom and stand on the floor where the work is done.
Bring the workflow that annoys you most. You leave with a different set of questions for your team, whether or not you ever hire me. If the answer is no, I will say so.
Rather write? guennael@celxadvisory.com