The AI-slop problem in IT marketing, and how to fix it.

More AI gave IT marketers more content. Buyers noticed. Here’s what changes when every draft starts from your proof.

CKCKFounder, Brand StoriOctober 7, 2026·6 min read
WHITE PAPER · INSURANCE · OCT 2026
Northwind
Claims in 2027.
Five shifts every claims leader should plan for now.
KEY FINDINGS
62%
of insurers will retire a claims system by 2027
2.1 pts
of combined ratio lost to leakage, on average
18 mo
median payback on claims data work
01Leakage becomes a data problem04
02Adjusters get a co-pilot09
03Fraud moves to real time15
04The core stops being the bottleneck21
05What to fund in the next 12 months27
PRIYA RAMAN · TOM HALE32 PAGES
PRPriya RamanHead of Data & AI, Northwind1 / 5
WHAT JUST HAPPENED
Open weights just closed the gap.
Here’s what it changes for regulated banks, and what it doesn’t.
Swipe
CUSTOMER STORY · CARD PAYMENTS
Northwind
4ms
Fraud decisions in four milliseconds, for 38 million cards.
Challenge
Rules caught fraud a day too late.
What we built
A real-time model on Databricks and AWS.
Result
$48M in fraud stopped in year one.
Three assets made from one company brain.

IT marketing teams were promised more content. They got it. A year later, the complaint from sales has changed: it used to be “this isn’t about our client”. Now it’s “this sounds like AI”.

Both complaints have the same cause. The writer, human or model, didn’t know enough about the buyer or the company to say something only you could say.

Slop is a context problem

A model with no context fills the gap with the most likely sentence. For IT services, the most likely sentence is the one every competitor already published: transformation, innovation, unlocking value.

More prompting doesn’t fix it. Each new chat starts from zero, so the writer retells the company story, pastes the case study and corrects the buyer’s priorities, every time.

“We say AI, and we show up like an old-school company.”— CMO, MID-SIZE IT SERVICES FIRM

Three tells a CIO spots in a paragraph

  1. No number. Real work has a result. “Faster claims” is a claim; “claims settled 40% faster” is proof.
  2. No buyer. The paragraph would read the same to a CFO and a CISO.
  3. No point of view. It agrees with everyone, so it says nothing.

Start from proof, not from a prompt

The fix is upstream of the writing. When every draft starts from what your company has actually done, for this buyer, the slop has nowhere to come from.

THE SAME SENTENCE, TWICE
FROM A BLANK PROMPT
We help insurers unlock transformative value with AI-driven innovation.
FROM YOUR COMPANY BRAIN · FOR AN INSURANCE CFO
Most leakage isn’t fraud. A top-10 insurer recovered $48M in a year by joining claims, policy and payment data first.

Notice what changed. Not the model, and not the prompt. The second sentence has a buyer, a number and a point of view, because the writer started from all three.

What to do this quarter

✓Collect your 10 best proof points in one place, with the numbers approved.
✓Write one point of view per industry you sell into. One page each.
✓Ban one phrase per month from every asset. Start with “unlock value”.
✓Measure content by meetings, not impressions.

None of this needs a bigger content team. It needs every writer to start from the same brain.

IT marketingAI contentThought leadership
CK
CK
Founder, Brand Stori · Host, CXO Spotlight

Spent his career in IT sales and marketing. Now building the company brain he wished he’d had.