Three Surprises From When We Embedded AI in Our Content Supply Chain
Jun 23 2026
By Corby Fine, VP Digital Marketing and Performance, IGM Financial
The fundamental question is how AI changes what your business is. Not how you do things. What WHAT you are.
I asked Claude directly once: could you be my financial advisor?
The answer was something like: “I can analyze 30 years of financial transaction history faster than any human possibly could. But I cannot read your body language when I give you the advice to see if it is going to scare the living daylights out of you. I cannot tell that your kid needs to save for school or that you are thinking of selling a house. That context matters.”
That answer tells you where the human element holds and where AI can run ahead.
“Governance might slow you down at first, but it actually makes the ambition sustainable. Don’t make my mistake.”
The Supply Chain We Had to Break
Our content supply chain had eight or nine steps. Brief, strategist, copywriter, designer, compliance loop, translation, build, push. Every handoff took time and was risky.
On March 25th we sent a newsletter to 400,000 people.
Article one: tax tips for singles.
Article two: mortgage refinancing.
I am not single and I am not refinancing. That newsletter was irrelevant to me and probably to most of our audience. That is what one-size-fits-all communications look like at scale.
Expensive to produce, largely ignored.
We are now saving about 50% of the time in that supply chain.
The brief is generated by an LLM from structured inputs. Translation went from third-party agencies to a self-serve model, our translators are now translation auditors.
75% of what used to go to our compliance team, our own compliance team confirmed we never needed to send them. Investment commentaries that used to take 48 to 72 hours now go out by lunch.
The Three Surprises
Surprise one: the technology was easy. The AI was ready in six months. The business took eighteen to catch up. Culture and governance in a traditional financial services organization are the hard parts. Not the models.
Surprise two: we got compliance wrong. We built first, proved value, then brought governance in. That sequence cost us months.
Surprise three: data quality is not a side problem. AI amplifies what you put in. The gaps and inconsistencies hiding in your systems become visible immediately, and a confident, fast, wrong answer is worse than no answer.
What AI Actually Takes
Is AI taking jobs? The reality is messier and more interesting. AI takes the parts of jobs that people did not want anyway.
We had a writer spending 60% of her time on production work. Formatting. Scheduling. Brief management. That is not why you hire a writer. She is now a brand curator.
The accountabilities shifted. The work that actually requires her got more of her.
The new roles that matter are the ones that did not exist two years ago.
Brand curators. Journey architects. AI adoption leads.
These are central to how enterprise functions in 2026. New jobs entirely.
“Don't waste talented people on mundane production work. Find better uses for them and automate the production.”
The Lens That Keeps You Honest
Before any AI initiative, the question gets personal and goes straight to WHAT this means for your business: “Is this a grow play, an optimize play, or a protect play?”
Know which one it is before you start. Most AI work right now sits in optimize. That is fine, but do not mistake efficiency for strategy.
Define what good looks like before you build. "AI will improve efficiency" is a wish.
A real target is specific: we will reduce this supply chain from nine steps to three, save 50% of the time, and redeploy those people to work that requires them.
If you cannot write it down before you start, you will not be able to evaluate it when you finish.
Architecture that is interoperable, composable, and open makes all of this faster.
The data surfaces better. The tools connect. The agents have something worth working with. The technology is ready. Is your company?