Inside the report
Both sides are betting on AI. Neither can see the whole deal.
Every retailer and CPG supplier is making the same bet this year: that AI will finally fix how trade gets planned, funded, and settled. Ninety percent expect their AI investment to keep growing in 2026, and nearly half are already piloting AI agents. The report’s argument is that none of it pays off without a precondition most organizations haven’t met — a shared view of the deal.
A supplier’s AI sees committed dollars against a plan. A retailer’s AI sees shelf performance against a calendar. Neither sees the other half, so neither can act on the whole deal — and on roughly $500 billion in annual trade spend, with up to 72% of promotions failing to break even, layering AI onto two disagreeing systems just automates the disagreement, faster.
It works because both sides are on it.
80%
reduction in joint business planning cycle time
35%
reduction in trade fund leakage
45 days
faster settlement
Early, customer-reported results from organizations running a connected model — the report walks through what’s behind each number.
In their words
The field is already saying it.
Ask retail and CPG trade professionals to describe their own process, in their own words, and the same gap surfaces again and again — independent of any vendor.
We have data analytics in place, but we are fully replacing it so that we not just look at past metrics but look forward… We will use AI to analyze these insights and predict behavior.
Describing their current trade platform
No AI or recommendation engine.
Describing their deal workflow
The approval process for promotional plans is complicated, leading to delays.
None of these are complaints about a product. They are descriptions of a process that was never built for two organizations to share — and they are why the shift underway is bigger than any single vendor’s roadmap.
What’s inside
The case for one shared view of every deal.
An agent trained on one side of a promotion can summarize that side. It cannot act on the whole deal. The report follows that precondition through the data — and through what early adopters of a shared model are already reporting.
$500B
planned and executed in trade spend every year
72%
of promotions never break even
58%
of retail & CPG orgs now actively deploying AI
The stakes, quantified: roughly $500 billion in annual trade spend, the up-to-72% of promotions that never break even, and the 1–3% of margin disconnected decisions leave behind.
Three converging forces: margin pressure, forward-looking planning, and agentic AI — why the shared-data problem is finally fixable in 2026.
Three shifts underway: from reporting to forecasting, from siloed execution to one commercial thread, and from dashboards to decisions.
The agenda for commercial leaders: five moves retail and CPG leaders can act on now — measured in the numbers a CFO actually cares about.
Quick answers
About the report.
- What is “AI Can Only See Half the Deal” about?
- The report argues that retail and CPG's AI investment in trade only pays off when both sides of a deal work from the same data. An AI agent trained on one side of a promotion can summarize that side — it cannot resolve a dispute, catch a mismatch, or act on the whole deal.
- Who produced the report?
- DemandTec, in partnership with Progressive Grocer and CGT (Consumer Goods Technology), published July 2026. It draws on research from DemandTec, Progressive Grocer, IDC, PwC Strategy&, Nielsen/McKinsey, and NVIDIA.
- What are the headline numbers?
- Retailers and their CPG partners plan and execute roughly $500 billion in trade spend every year; up to 72% of trade promotions never break even; 58% of retail and CPG organizations are now actively deploying AI; and disconnected pricing and promotion decisions cost an estimated 1–3% of margin in year one.
- How do I get the report?
- Fill in the short form on this page and the PDF opens immediately. It is independent research and field data — no sales call required.
See both halves of the deal.
The bilateral demo shows the retailer and CPG experience side by side, on one shared source of truth.

