People-First: Investing most in what AI can't automate

Last week, we started to explore the success factors of retail’s quiet AI winners. These winners share three characteristics: focusing initially on operational improvement, prioritising change management, and proving value before scaling. Today, we look harder at the second characteristic; retailers quietly succeeding with their AI implementations are investing around 70% of their resources not in technology, but in their people.

Challenging the Media Narrative

In this new age of AI, the future of work is often portrayed by the media as a zero-sum game between a human workforce and the machines. BCG’s research reveals a different reality. Companies successfully implementing AI allocate 70% of transformation resources to people and processes, 20% to data infrastructure, and just 10% to algorithms.

The evidence is very interesting. Companies succeeding with AI are deeply invested in the change management required to reorganise how their people work and the engagement required to bring them and the organisation, on the transformation journey. They are doubling down on their human capital.

Global Patterns of Success

Walmart committed $1 billion to skills training through 2026, creating AI certification programs with OpenAI. Chief People Officer Donna Morris says: “We’re changing the ratio of jobs to technology, not the number of jobs.”

The investment includes AI literacy programs, prompt engineering training, and new role creation. Store associates use AI for inventory insights while maintaining personal customer relationships. The technology amplifies capability, it doesn’t eliminate jobs, rather it frees specialists for the complexities of personal service requiring human judgment.

Decathlon is an excellent case study of people-first transformation at scale. The company focused on empowering their store teams across 2,080+ outlets, giving them AI-generated insights showing customer patterns to enable them to better serve (and sell to) their customers, in store, every day. Crucially, Decathlon backed this up with heavy investment in training these teams to interpret and act on insights. This store and department-level CRM means staff have informed conversations, with AI providing the intelligence and people providing that vital service connection.

Myer’s six-year transformation prioritised employee engagement through their M-Metrics platform, connecting 10,000+ team members with real-time insights. They equipped staff with 3,600 devices and created four applications: stocktake (reducing inventory discrepancies 15%), price marking (enabling same-day promotional changes) and receiving/dispatch (cutting processing time 30%). Result: 20% faster transactions and satisfaction jumping from 65% to 85%, through augmented human service, not automation.

Nike’s distribution centres have taken this a step further. They didn’t simply install 1,000+ robots into their distribution centre processes. This technology fundamentally changed the system of work within these DCs, so Nike invested extensively in re-training workers to operate within these new systems. As a result, new roles emerged, such as robot coordinators, automation specialists, and human-robot interaction designers. Productivity soared and employment actually increased.

The Cost of Getting It Wrong

McDonald’s AI drive-thru was a partnership with IBM, designed to automate order-taking, reduce labour costs and speed up service in 100 test locations. The experiment was fraught with technology issues. The AI struggled with accents, background noise and menu modifications. The system couldn’t handle common requests like “no ice” or “extra sauce.” Often, AI order-taking slowed the whole process down due to repeated corrections.

But the failure went beyond technology. Employees weren’t trained to monitor and intervene when AI failed. Customers received no education on how to interact with the fallible technology. Staff morale dipped markedly as they became the target for customers’ frustrations with the technology and the waiting time. McDonald’s focused on eliminating order-taking staff rather than empowering them to work alongside AI or prepare customers for the change.

The result was both technological and organisational failure. There was a fundamental misunderstanding of the value of human service staff to the efficiency of the drive-thru workflow and customer experience. In June 2024, McDonald’s ended the IBM partnership and removed the system from all test locations. A cautionary tale about deploying technology without investing in the people who make it work.

Success Factors for Transformation

Winners share consistent practices:

Early stakeholder engagement at every organisational level, top to bottom.

A clear and compelling vision of the role AI will play in the future of the business and the working lives of its people.

Workflow redesign that reimagines existing processes rather than replacing them wholesale.

Cross-functional teams that bring together technical, operations and frontline staff to co-create solutions.

Pilot programs with feedback loops, allowing employees to shape AI integration

Reimagining work systems, driving new role creation and structured re-skilling programs with clear career pathways.

Incentive realignment that rewards human-AI collaboration, not just efficiency metrics.

Sustained financial and engagement investment, maintaining the 70% throughout transformation.

Common Pitfalls to Avoid

Failures follow predictable patterns:

Seeing AI as a path to cost reduction and measuring only efficiency gains

Procuring shiny solutions without clear use cases

Creating adversarial “humans versus machines” narratives

Creating “AI silos” where only certain teams are enabled

Seeing this change as an HR problem

Offering no clear, motivating vision of the future of teams’ work

The Choice Ahead

The next 12-18 months will widen the gap between retailers who invest in people and those who don’t. Winners will announce improved margins alongside higher employee satisfaction. They’ll reimagine whole systems of work, re-bundling tasks and creating new roles that haven’t existed before. They’ll continue to invest in building human-led competitive advantages that their competitors can’t replicate with technology alone.

The question is whether you are still looking to invest in technology to replace people, or to make them irreplaceable. The quiet winners have already chosen.