Frontier Firms: An organisational model for sustainable AI value creation

Artificial intelligence (AI) has long been a recurring strategic topic on executive boards, supervisory boards and management committees. Investors expect tangible progress, customers demand high-quality digital experiences, and competitors are increasingly using AI as a differentiating factor. Moreover, uncertainty remains high: where does real added value lie? Which investments pay off? And how can productivity potential be scaled without increasing governance or compliance risks?

In many companies, a high willingness to adopt AI is offset by limited actual impact. Pilot projects may demonstrate isolated successes, but they lose momentum as soon as scalable structures are lacking. Successful organisations, on the other hand, do not view AI as a stand-alone technology, but as an integral part of their operating model.

Frontier Firms as an organisational model Frontier Firms – those pioneers that Microsoft describes as “human-led, AI agent-operated” – follow an organisational principle in which human judgement and AI-supported process execution are clearly distinguished from one another. People set goals, make decisions and bear responsibility. AI agents analyse data, orchestrate workflows and execute multi-stage processes across systems. This interplay gives rise to a new working model:

task_alt

Technology handles operational tasks and overall coordination

task_alt

Staff focus on assessment, prioritisation and accountability

Frontier firms are distinguished not by their use of individual tools, but by a fundamental rethinking of their value creation logic and operational mechanisms. This model is not only aimed at digital pioneers, but also offers a practical approach for companies seeking to adopt AI in a deliberate, scalable and responsible manner.

Value-driven approach rather than a focus on technology

In many organisations, there is currently a great deal of activity surrounding AI, whilst the actual value created often remains limited. Initiatives fail less because of technological constraints than due to a lack of value logic and scalability.

Frontier firms align AI initiatives with clearly defined use cases that contribute to business outcomes, cost efficiency or risk reduction. The expected value is quantified, suitable KPIs are defined, and only then is a technology decision made.

One example of this is Vodafone: following a successful pilot, the roles and processes in which AI offers the greatest added value were identified. On this basis, Microsoft Copilot was rolled out to 68,000 employees.

Work processes and role allocation Once the value has been defined, the question arises as to how work processes must be designed to make the expected benefits achievable. Without adapting roles, responsibilities and workflows, AI often remains isolated.

Frontier firms therefore view AI as an integral part of work design rather than merely a technology implementation. AI agents take over analysis, coordination and routine processes across system boundaries. Employees focus on decision-making and accountability for results. Customer projects also follow this pattern: at the Australian telecommunications provider Telstra, an AI-based compliance process with over 20,000 risk-based controls was developed. This enabled the team to concentrate on high-risk decisions – a clear productivity gain and a contribution to greater strategic control reliability.

Governance as a prerequisite for scalingProductivity gains can only be scaled if there is trust in the use of AI. This requires transparency, reliability and security – anchored in a robust governance model. If governance is only considered at a later stage, this creates uncertainty among managers and employees.

Frontier firms therefore integrate governance at an early stage and define framework conditions across the entire AI lifecycle. The aim is not greater control, but a clear regulatory framework that empowers teams to tackle new use cases safely and confidently

Ready to talk? Contact us

Specific recommendations for managers

Based on the patterns described for frontier firms, three clear recommendations for action can be derived:

Always prioritise values over technology

Focus on a small number of use cases with clearly measurable business benefits. Define the expected value and appropriate KPIs before making any technology decisions. Projects without clearly defined benefits generally cannot be scaled effectively
Redesign work processes in a targeted manner

Think of AI as an integral part of how work is organised. Clearly define which tasks, decisions and responsibilities remain with staff, and where AI agents support or take over processes. This is the only way to achieve lasting acceptance and impact in day-to-day operations.
Integrate governance at an early stage

Establish clear guidelines for the responsible use of AI right from the start. A clearly defined governance framework builds trust, reduces uncertainty and lays the foundation for scalable productivity gains.