3 actionable ideas in 3 minutes, only for Retail Leaders
Welcome to DX Brief - Retail, where every week we review industry podcasts and reports to share what’s insightful and what you can do about it.
1. 45M Ways to Burn Money on ERP (and How Retailers Can Avoid It)
The Catch Up Podcast, Ep: 45 Million Reasons to Get Your ERP Strategy Right (Sept 25, 2025)
For: CIOs, CFOs, Program Directors, SIs and PMOs.
Key takeaway: If you outsource ownership to your SI, skip test and data leads, and under-spec the RFP, you will pay for it. In one case, literally £45M with nothing to show.
The failure pattern: Enterprises start large D365 or SAP programs with no client-side ERP experience in streams, treat the SI as the source of truth, and accept upsells they don’t always need. Teams skip dedicated testing and data migration leads, let weak designs through, and only discover gaps after go-live is already late. Result: multi-year spend and canceled modules. This is not rare. Surveys and case logs show high ERP failure rates and well-known blow-ups across industries. (CIO)
What to do before kickoff
Treat the RFP as a product spec: Invest weeks, not days. Capture tricky realities like multiple payrolls, project accounting, and reporting constraints. Bring an experienced program lead in at RFP time.
Staff your side: program manager, workstream leads, test lead, data migration lead, and a strong PMO. Even with a top SI, you must own critical decisions.
Demand a real integrated plan: Building and reviewing a credible plan takes 3-8 weeks depending on scope. Rushing this step is how budgets die.
Guardrails during delivery
Separate “musts” from “nice-to-haves.” Resist SI upsells that aren’t tied to measurable value.
Lock a defect-driven test strategy and data quality gates early.
Insist on process fit. ERP failures often trace back to poor mapping of legacy workflows to the new system.
2. Stop Patching Legacy Systems. It’s Time to Design Your Stack Around the Customer
The RetailWire Podcast, Ep 47: Navigating Retail’s Digital Future (Sept 25, 2025)
For: CIOs, Heads of Omnichannel, Merch Ops, Supply Chain, SI leaders.
Key takeaway: Your bottleneck is not “AI.” It is legacy integration and fragmented data. Treat architecture as strategy, not a toolbox.
Why it matters: Most retailers still run on old ERP, POS, and supply chain systems that were never built for real-time omnichannel. Layering cloud and AI on top without unifying data creates fragmentation, security gaps, and rising cost. Build a unified architecture that connects product lifecycle data to customer data. Design for scalability and interoperability from day one.
The impact of new systems: Case studies show how new ERP+BI cut reporting from weeks to hours and reduced project overruns by 30% (SC&H). RFID programs lifted stock accuracy from ~70% to ~98%, raising service levels from 85% to 98% and cutting waste from 5% to 2% (RFID Journal).
Trend to watch: AI-driven personalization
Retailers are moving from isolated tactics to end-to-end AI personalization across marketing, merchandising, and service. Industry outlooks and NRF 2025 coverage stress AI agents, CDPs, and privacy-by-design as core to growth. McKinsey calls AI-powered personalization an increasingly proven driver of revenue and efficiency; NRF flags 2025 as the year of the AI agent; Gartner reports 85% of service leaders plan to pilot customer-facing genAI in 2025.
Case in point: Saks Global reports AI-personalized homepages lifted revenue/visitor by 7% and conversions by ~10% in early tests. Vogue Business
What to do this quarter
Map the data plane: where SKU, content, price, inventory, and customer IDs live; define your system of record per domain.
Stand up a real-time inventory backbone and item-level tracking for your top categories. Pilot RFID where returns and stockouts hurt. (ecrloss.com)
Replace “projects” with an experimentation cadence: pilot, measure, scale. Prioritize quick wins that prove value but design for long-term scale.
3. Retail AI That Pays For Itself: From Hype to Human-in-the-Loop
The Enterprise Edge, Tim Sears, Chief AI Officer, HTEC (Sept 26, 2025)
For: CAIOs, CIOs, Data leaders, Retail Tech vendors.
Key takeaway: Retail AI tech is real, but most companies are bad at deploying it. Kill dabbling. Pick use cases with true ROI and keep a human in the loop as you scale toward agents.
Why many AI projects stall
Executives rush into chatbots and pilots without process redesign, org changes, or value math. Industry surveys show adoption rising but very few firms call their genAI strategy “mature,” and reported gains are often small single-digit savings (McKinsey & Company). Some headlines even claim most companies see no ROI, which should reinforce the need for disciplined selection and measurement (The Times of India).
How to pick winners
Sears’ rule: ROI is arithmetic. Map expense pools, revenue levers, and missed opportunities; rank use cases by payback, not coolness. Filter anything that does not self-fund. The era of dabbling is over.
Build for impact, not demo
Human-in-the-loop first. Get review and escalation right before agents execute workflows. Treat semi-autonomy like self-driving cars; it will take longer than the hype says.
Teach the enterprise. Broad internal training accelerates adoption and changes how engineers work.
Start where data is clean and impact is obvious: service ops, supply chain exceptions, and content ops are common early winners. External surveys show cost benefits concentrate in these areas. (Stanford HAI)
Getting real about agents
Agentic AI is moving from chat to work execution, but maturity lags and most enterprises are still learning. Pilot agents in bounded flows with auditable logs and SLA metrics. (EY)
What to do now
Define 3 ROI-positive use cases with clear owners and SLAs.
Stand up feedback capture and model improvement loops.
Keep the kill switch close. If it does not pay, pause it, fix data and process, then resume.
Disclaimer
This newsletter is for informational purposes only and summarizes public sources and podcast discussions at a high level. It is not legal, financial, tax, security, or implementation advice, and it does not endorse any product, vendor, or approach. Retail environments, laws, and technologies change quickly; details may be incomplete or out of date. Always validate requirements, security, data protection, labor, and accessibility implications for your organization, and consult qualified advisors before making decisions or changes. All trademarks and brands are the property of their respective owners.