At Microsoft, Dhaval Desai has spent the past several years confronting a challenge increasingly familiar to global executives: how do you transform an operation that cannot afford to slow down? His answer doesn’t begin with a technology, but with a more enduring discipline of simplification, strong data foundations and better decision-making.

For a company operating at Microsoft’s scale, supply chain is far removed from the conventional image of warehouses, freight and inventory. It stretches from consumer devices such as Xbox and Surface to the server racks and hardware infrastructure underpinning Azure’s global cloud operations.

Managing that landscape not only requires a traditional supply chain discipline, but it also demands something more, the ability to transform continuously while the business continues to grow.

That is the environment in which Dhaval Desai, Partner Group Engineering Manager, Supply Chain at Microsoft, operates. Desai joined Microsoft in 2021 after working across supply chain roles at Verizon and SAP, a career path he describes as being connected by one consistent idea: using technology to solve increasingly complex business problems.

At Verizon, that meant learning how operational decisions translate into business outcomes. At SAP, it meant helping organizations across industries and geographies transform planning and execution processes. Microsoft offered a different proposition. Instead of enabling transformation from the outside, Desai could participate directly in driving it inside one of the world’s largest technology companies.

That distinction has proved significant.

Two Supply Chains, One Set Of Fundamentals

Microsoft’s Devices and Cloud businesses illustrate just how differently the same supply chain principles can manifest themselves.

Devices are product-centric. Hence, planners must interpret consumer demand, product launches, seasonality, channel dynamics and relatively short product life cycles. Cloud infrastructure introduces a different form of complexity. Microsoft must plan the physical servers, racks, networking equipment, processors, memory and other hardware that ultimately supports services consumed through its data centers.

Those investments also operate over longer horizons. Cloud planning must account not only for computing demand, but for capacity, capital deployment and physical constraints such as energy availability. Errors that appear small in percentage terms can become substantial when applied at enormous scale.

Yet Desai argues that the underlying disciplines remain remarkably consistent.

“Technology changes; fundamentals do not,” he says in his written responses.

Demand planning, supply planning, inventory optimization, supplier management and execution excellence remain important regardless of the product being delivered. What changes is the context in which those principles are applied.

That distinction has shaped Desai’s approach to transformation. The objective is not to discard established supply chain thinking every time a new technology arrives. It is to create an operating model capable of applying those fundamentals faster, more consistently and at greater scale.

Transforming While “The Airplane Is Still Flying”

Desai has a useful description for transformation at a rapidly expanding business: transforming while the airplane is still flying.

Traditional transformation programs can implicitly assume a degree of organizational stability. There is a current state, a desired future state and a period in which the company moves between the two.

Hypergrowth undermines that assumption.

Operations have to scale while systems are being modernized and new capabilities are introduced. Customer requirements do not pause while transformation teams redesign processes or implement new platforms.

The result, Desai argues, is that companies should prioritize adaptable processes over perfect processes.

Complexity that should have been removed does not suddenly become useful because it has been digitized. Automating an inefficient process can simply produce inefficiency faster. Desai therefore begins with what he calls a “lean first” approach, removing unnecessary complexity before introducing technology.

Data comes next because planning, supplier engagement, logistics and execution cannot operate coherently if each depends on a different version of reality.

Only then does modernization and automation begin to deliver its full value.

The final component is explicitly human. Systems may execute processes, but the business still has to own them. Change management, user adoption and clarity of accountability determine whether transformation actually takes hold.

AI Changes The Speed, Not The Fundamentals

That philosophy also explains Desai’s view of generative AI.

For many businesses, the AI conversation begins with the technology itself. Which models should we use? Should we build an internal AI team? Which functions can we automate?

Desai reverses that sequence.

The meaningful transition, in his view, occurred when AI stopped being primarily an impressive demonstration and began solving identifiable business problems. Better foundation models, copilots, enterprise AI platforms and increasingly mature cloud data environments have made those applications easier to build and deploy.

But the business case still needs to come first.

“Don’t ask where AI fits,” he advises. “Ask where people spend time on low-value work that technology can augment.”

His first step for an organization considering an AI transformation is deliberately practical: identify one meaningful business problem.

From there, companies should define measurable outcomes and determine whether the underlying data and processes are mature enough to support the intended use case. Technology selection comes later.

The opposite approach creates predictable problems. Desai identifies three in particular: starting with technology rather than business value, creating AI strategies disconnected from operational priorities, and assuming AI can compensate for poor data or broken processes.

It cannot.

In supply chain planning, AI can reduce the amount of time planners spend searching for and assembling information. It can surface risks, identify exceptions, provide recommendations and expose both strengths and weaknesses in existing processes.

For executives in smaller organizations, that leads to another lesson: resist copying the architecture of a company whose scale bears little resemblance to your own.

Microsoft may require sophisticated engineering organizations, complex optimization capabilities and large platform investments. Many companies do not. Desai’s recommendation is to use modern cloud and AI platforms where possible rather than automatically assuming capabilities need to be built internally.

His formulation is simple: “Scale changes the tools you need, not the fundamentals you should follow.”

Clarity As A Leadership Product

Technology, however, is only part of Desai’s transformation thesis.

Rapid growth tends to push leaders toward structure: more processes, clearer reporting lines and greater control.

Desai believes the more important requirement is alignment.

People need to understand the mission, the priorities and the framework within which they can make decisions. Once those elements are clear, empowerment can scale more effectively than centralized control.

That becomes particularly important during periods of uncertainty, when Desai argues leaders should communicate more frequently.

“As organizations scale, clarity becomes a leader’s most important product,” he says.

Culture follows the same logic. Employees need the resources and freedom to develop new capabilities as technology changes around them. Continuous learning cannot simply be treated as an HR initiative. It becomes part of the operating model itself.

That may ultimately be Desai’s most consequential argument about AI.

The objective should not be to replace expertise, but to increase its leverage.

“Use AI to augment expertise, not replace expertise.”

For all the excitement surrounding generative AI, the technology itself may not determine which organizations extract the greatest value from it. Desai’s argument is that the stronger position comes from combining sound operating fundamentals, high-quality data, empowered people and AI-enabled decision-making.

It is a fitting conclusion to a career that has moved through ERP systems, cloud platforms and now generative AI.

The technology has changed dramatically.

The management challenge has not.

The work is still about helping organizations make better decisions, make them faster and make them at scale.