Microsoft has formed a consulting practice that will help enterprises build AI software.

The business unit, which is called the Microsoft Frontier Company, is launching with a $2.5 billion budget. The software giant didn’t specify how the capital will be spent. Much of the sum will presumably go towards payroll expenses: Microsoft has assigned 6,000 employees to the unit. There are also other costs associated with AI projects, notably the hardware required to fine-tune LLMs.

Many Microsoft Frontier Company staffers are forward deployed engineers. Those are developers who work at client organizations’ offices. In high-stakes software projects with a large number of stakeholders, an on-site presence helps to build rapport that can go a long way towards speeding up development. It also eases certain cybersecurity tasks.

Quite a few of Microsoft’s consulting partners offer forward-deployed engineering services. We asked one such partner, Insight Enterprises, about the impact of the software giant’s move into the market. 

“Microsoft is validating our core viewpoint: that the tech is great, but enterprises need a hand to get to best value,” Miles Ward, Insight’s CTO for AI, told Boardroom Insight. “Microsoft is big, but the market of customers who need help is much, much bigger.”

The executive pointed to established consulting firms’ client relationships and and multi-vendor focus as differentiators. 

“A Microsoft FDE is going to do amazing things with Microsoft tools. Insight FDE’s can help you across your whole technology estate, creating integrated agentic systems that meet your users and customers where they are, and take them to what’s next,” Ward said. “Hyperscaler FDE departments are months old: Insight has been working with some of our customers for decades. There is the domain expertise, familiarity with internal norms and nomenclature – it is the depth of experience with our customers’ nuances that sets us apart.”

The Microsoft Frontier Company’s work will encompass not only LLM development but also certain related tasks. Notably, the software giant stated that the unit can help clients secure their AI applications. That will presumably involve building guardrails against AI-specific hacking tactics such as prompt injections, malicious prompts hidden in web pages and files. 

Microsoft also listed governance and observability among its new consulting practice’s focus areas. Governance is a term that mainly refers to regulatory compliance. Observability, in turn, is the ability to monitor that a service works as intended. In the case of AI applications, that means tracking the response quality of the underlying LLMs.

Photo courtesy of Microsoft