MACH X 2026: Spotlight on agent-ready retail

Retailers are finding that AI deployments and scale depend less on models than on disciplined data, architecture and governance, Miya Knights, Retail Technology Publisher, reports

Retailers attending MACH X in Toronto this week heard that artificial intelligence (AI) is pushing composable technology from architecture principle to business necessity.

The event, hosted by the nonprofit organisation that advocates for a composable approach to IT architecture and integration (using microservices, application programming interface (API)-first, cloud-native and headless, i.e, MACH), brought together end users, vendors and systems integrators (SIs) to consider how enterprise technology must change as agentic AI moves from experimentation into production.

For the MACH Alliance, the shift is also forcing a refinement of its own message. Jason Cottrell, chief executive and founder of Canadian SI Orium and MACH Alliance president, said the now five-year-old organisation was maturing beyond foundational MACH technologies, while keeping its original principles intact.

“We’re focusing on MACH, the concept, so that’s speed, performance and scale,” he told Retail Technology in an exclusive interview. “Microservices, API-first, cloud-native and headless are all still parts of what we’re looking at, but we need the flexibility to recognise that API-first may not be as relevant when it’s agents connecting agent to agent, even though most enterprises are currently only doing agent to system deployments.”

Getting ready for agents

Cottrell said the Alliance’s new “agent-ready” certification was designed to help enterprise technology buyers cut through market noise, while recognising that most enterprises are still at a foundation-building stage.

“[This certification] is about confirming real use cases deployed against products,” he said. “We can certify the foundations: how you have systems that agents can work with, how you get your data ready, and how you get logic and tooling ready, open and available to agents.”

His comments reflected the wider tone of the event. MACH X was not a showcase for AI hype. Discussions focused on why many enterprises, including retailers, are still struggling to turn AI investment into measurable business outcomes.

Defining value before data

Loblaw Digital senior director of strategic partnerships and AI strategy, Chrissy Munroe, told attendees that value definition had to precede build-or-buy decisions.

“If you don’t define what success looks like upfront, you end up with a roadmap that is just a bunch of demos,” she said. “The discipline is what keeps technical work connected to business reality.”

Munroe said the digital arm of Canada’s largest food retailer had built around 45 custom AI solutions over four years. 

But she also stressed that the business does not treat all value in the same way. Some use cases are measured by revenue, margin or cost savings. Others create operational leverage through speed, quality or accuracy. A third group supports longer-term strategic impact.

Optimise workflows first

“You don’t need to force every AI initiative into the same return on investment template,” added Munroe. “But you do need to be explicit about what success looks like.”

Loblaw’s examples included a space-planning tool that supported the expansion of its No Frills discount banner, using data such as store plans, category contexts and local demand signals to support AI-assisted allocation decisions. Munroe said the tool had become effective enough for the business to apply its recommendations to more of the existing store network.

But Munroe’s strongest warning was against using AI to accelerate badly designed work.

“Do not automate broken workflows,” she said. “Go slow enough to ask what the work should be, so you can go fast to deliver value.”

Developing AI differentiation

For ALDO Group, the most important distinction is between AI access and AI advantage. Matthieu Houle, CIO at Canadian multinational shoes and accessories retailer ALDO Group, said giving employees tools such as ChatGPT or Copilot was increasingly a basic requirement, not a differentiator.

“That will not give you any differentiation against the competition,” he said. “It is the new kind of Office tool. We are all more productive, but you are not going to compete on that.”

Houle said the bigger opportunity lies in choosing meaningful agentic AI use cases that can change core retail economics. At ALDO, that means planning. The footwear retailer has invested in an agentic retail planning startup after realising that legacy planning processes were no longer suited to faster tariff changes, supply chain volatility and seasonal unpredictability.

“The front end was pretty modern,” said Houle. “But the retail business was trying to follow tariff changes and supply chain problems with a process that used to run twice a year. Now they have to do this almost every two weeks.”

He added that the early work had changed his expectations of what could be built and how quickly.

“The amount of work and the things we can build in four months is something I’ve never seen in my career,” he said. “It changes the process internally. It’s changing the core.”

Creating a data strategy

Holt Renfrew senior vice president of information technology, Alicia Samuel, offered a counterpoint to the pace of ALDO’s agentic planning development: she stressed that retailers cannot move fast with AI unless their foundations are secure enough to support it.

Samuel said the Canadian luxury department store chain had spent recent years modernising its IT infrastructure, developing a data strategy, establishing an AI committee and policies, and building an AI-ready architecture with technology partners Conscia and Amazon Web Services (AWS).

“Everybody wants to push forward on AI,” she said. “But we had data that we needed to take care of foundationally, because that is our gold. 

“We needed to make sure we were doing all the right things because we don’t know what we don’t know. The minute we unleash such a strong technology, we will quickly see what we don’t know.”

Single source of truth

Samuel said the key advice she would give other retailers was to “get your data right” and build a single source of truth.

“You need to have the data foundation appropriately done, because it is garbage in, garbage out,” she said. 

“When you start executing on agents and agent-to-agent, without a human loop, you have now unlocked possible issues if you do not have a proper data governance, security and loss prevention strategy in place.”

She also stressed the importance of a semantic layer, describing it as the business translation that ensures revenue, profit and sales have consistent meaning across the organisation, irrespective of the data used to measure them.

Establishing firm foundations

Taken together, the retailer presentations suggested that AI is giving MACH new urgency. Composable technology is no longer only about faster digital change. 

It is becoming the operating foundation for agentic systems that require trusted data, governed access, reusable capabilities and clear accountability to deliver greater efficiencies and productivity.

Alliance president Cottrell said the agentic market was still too early for definitive standards across all protocols and control points. But he argued that the Alliance was already best positioned to define what readiness looks like.

“You don’t want to programme an agent and just say, ‘Please only do these things,’” he said. “You want to set intent in the agents, but you also want to enforce it in the key systems of action.”

Celebrating AI success

The event closed with the Alliance’s fifth annual Impact Awards, recognising measurable progress in open, composable and connected IT architecture and including its first Agentic Achievement cohort for autonomous AI systems in live production. 

Winners spanned grocery, automotive, fitness, fashion and other sectors, with UK retailer Frasers Group named among the Composable Impact cohort for its work with Alliance members, content management system provider Amplience and digital commerce agency Lab Digital, on a more flexible, composable commerce platform for Sports Direct and the wider group.

For retailers, the message from MACH X 2026 was clear. The winners will not be those with the most AI pilots. They will be those who can connect AI ambition to business value, data discipline and composable architectures built for scale.