Published for RETHINK Retail, where the world’s most consequential retail leaders come to think clearly, act decisively, and stay ahead of what is coming next.

Most retailers have run a connected store pilot. Fewer have successfully scaled one. That gap from an inspiring proof-of-concept to an operational solution working successfully at hundreds of sites is where the connected store dreams falter most often, and this is the key issue for retail technology in 2026.

The data is sobering. IBM’s 2025 global CEO study, referenced at Google Cloud Next 2026, found that only 25% of AI initiatives delivered expected ROI and just 16% scaled enterprise-wide. ¹ 

That is not a technology failure. The reason for this phenomenon is a lack of governance, a platform, and readiness to operate; this is where the RETHINK Retail professional network operates daily.

With its seminars, analysis, and intelligence, RETHINK Retail helps retail executives get an honest background on how pilots fail, what scaling really takes, and what separates the successful retailers from everyone else.

Why Pilots Stall and Scale Does Not Follow Automatically

A pilot requires a completely different set of circumstances for success than the rollout does. The pilot is conducted within a controlled setting with specific resources and even a vendor team on hand. A rollout runs in the real world, with stretched IT teams, store staff who were not involved in the design, and infrastructure that was never built for the load.

IBM’s Think 2026 conference in May 2026 addressed this directly. The organizations that are moving forward have not increased the amount of AI in their organization but have redesigned their business models,” explained the Chairman and CEO of IBM, Arvind Krishna. As per data from IBM, only about 25% of all AI projects provide expected ROI because of problems with a lack of policies, ineffective orchestration, and non-integration with core business operations. ²

A June 2025 global survey conducted by IBM among 2,900 executives revealed that 83% believed that AI agents would increase process efficiency in 2026. Further, 64% of AI funds were being allocated to core business operations compared to past standalone projects. The shift from pilot thinking to platform thinking is not just a strategy preference. It is where the returns actually live. ³

Microsoft: The Organisations That Scale Have Redesigned How Work Gets Done

At NRF 2026, Microsoft was unambiguous about what separates retailers who scale from retailers who stall. The organisations winning are those building a unified, intelligence-driven operating model, not adding connected store technology on top of existing fragmented workflows.

Microsoft’s Forrester Total Economic Impact study (May 2026) documents what that approach delivers in practice: $14M to $23.9M in total three-year benefits across marketing, supply chain, and store operations for retailers that scale these use cases, with 124% to 282% ROI over three years. The study is explicit that most organisations will capture the first wave of gains and then stall, with the gap between those who compound their advantage and those who plateau determined entirely by whether the organisation redesigned incentives, workflows, and management systems to support scale.

That redesign is exactly the kind of operational conversation RETHINK Retail’s community is built for. Not the technology decision in isolation, but the organisational model that makes the technology perform at scale.

Google Cloud: The Experimental Phase Is Behind the Leaders

At Google Cloud Next 2026, CEO Thomas Kurian opened with a statement that defined the industry moment: “You have moved beyond the pilot. The experimental phase is behind us. And now the real challenge begins: how do you move AI into production across your entire enterprise?”

The retailers who have answered that question on Google Cloud’s platform are already visible. Walmart has deployed Gemini Enterprise internally to help store and supply chain team leaders connect to business data more quickly to better serve customers. The Home Depot, building on a 10-year Google Cloud partnership, is using Gemini Enterprise to deliver product expertise to every customer on every channel at any hour. These are not pilots. They are production systems at enterprise scale.

Google Cloud’s 2026 AI Agent Trends report for retail, based on 585 senior leaders from retail and CPG organisations, confirms the direction: the industry is moving beyond simple automation to intelligent agentic systems that streamline workflows, secure operations, and engage shoppers across every touchpoint. The retailers building that architecture now are the ones who will define what connected store leadership looks like in 2028.

Cisco: Scaling Requires a Network Built for the Load

Every connected store that successfully scales does so on a network infrastructure that was designed for scale from the beginning. This is not a detail that can be retrofitted after deployment.

Cisco’s State of Wireless 2026 report, surveying 6,098 wireless decision-makers across 30 markets (April 2026), found that businesses investing holistically in wireless alongside AI, automation, and security achieve 63% higher average ROI. The same report found that 98% of those using AI automation report substantial gains, saving an average of 3 hours and 20 minutes per person per day.

The challenge is real. 97% of retail organisations report rising operational complexity driven by IoT, IT, and OT workloads, and 39% report bandwidth challenges tied to video analytics, digital signage, and streaming content. ¹⁰ 

Retailers scaling connected stores are not just adding devices. They are adding intelligence, real-time data streams, and autonomous workflows to a network that must carry all of it without degradation. The retailers who treat network investment as a strategic decision rather than an infrastructure cost are the ones who scale without hitting the ceiling.

Palo Alto Networks: Security Gaps Are What End Scale Programmes

The moment a connected store programme moves from pilot to rollout, the attack surface grows in direct proportion to the deployment. Each new location, each new device, each new integration point is an opportunity for a threat actor to find a gap.

Palo Alto Networks’ Unit 42 Global Incident Response Report 2026, drawn from over 750 major incidents across 50 countries between October 2024 and September 2025, found that attacks are now 4x faster, with the fastest intrusions reaching data exfiltration in just 72 minutes. Critically, in more than 90% of incidents, preventable gaps materially enabled the intrusion, including inconsistent control deployment, incomplete telemetry, and over-permissive identity trust. ¹¹

For retail specifically, the Unit 42 report identifies data theft and extortion tactics targeting retail and hospitality organisations as a notable and growing shift in the cybercrime landscape from mid-2025 through early 2026. Security architecture is not a post-scale consideration. It is the foundation that determines whether scale is achievable in the first place. ¹²

From Pilot to Production: What It Actually Takes

Scaling a connected store is not a technology problem. It is a leadership problem. The organisations doing it successfully in 2026 share four characteristics: they have rebuilt their operating model around the platform, not the pilot; they have invested in network infrastructure as a strategic asset; they have embedded security from the first device; and they have built the internal governance to sustain what the technology makes possible.

RETHINK Retail exists to give retail leaders the peer-level intelligence to make those decisions with confidence, from practitioners who have made them before, and with the honest account of what worked and what did not.

The gap between retailers running connected stores at scale and those still stuck in pilot purgatory is widening every quarter. The conversation that closes it starts here.

Explore more at RETHINK Retail: https://rethink.industries

References

  1. BizTech Magazine — Google Cloud Next 2026: Expanding AI Agent Adoption Requires Culture Shift — 24 April 2026
  2. IBM — 2026 Goals for AI and Technology Leaders — 14 January 2026
  3. IBM Newsroom — IBM Study: Businesses View AI Agents as Essential, Not Just Experimental — 10 June 2025
  4. Microsoft — Microsoft Propels Retail Forward with Agentic AI Capabilities — 8 January 2026
  5. Microsoft Cloud Blog — Agentic AI Is Reshaping Retail and Consumer Goods Economics — 21 May 2026
  6. BizTech Magazine — Google Cloud Next 2026: Businesses Are Moving Into the Agentic Era — 23 April 2026
  7. Google Cloud — Next 26: Building the Agentic Enterprise — May 2026
  8. Google Cloud — AI Agent Trends in Retail and CPG 2026
  9. Cisco — State of Wireless Report 2026 — 2 April 2026
  10. Cisco — 5 Wireless Trends Retail IT Teams Can’t Ignore in 2026 — 15 April 2026
  11. Palo Alto Networks — 2026 Unit 42 Global Incident Response Report — 17 February 2026

 RH-ISAC — 2026 Unit 42 Global Incident Response Report — March 2026



🔒 Login or Register to continue reading