Many companies struggle to adopt Generative AI in a way that delivers real value. Fujitsu tackled this challenge using platform thinking, enabling employees to create custom GPTs tailored to their needs. By using AI as a trusted assistant, the company turned GenAI into an internal knowledge-sharing platform, bridging teams and enhancing business processes effectiveness. Their approach highlights three key principles: enable, don’t impose; customers, not suppliers; and agile, not stage-gate. Embracing these principles, organizations can move beyond pilot projects and drive meaningful, scalable AI adoption.

Rethinking GenAI Adoption: Why Traditional Approaches Fall Short

The adoption of Generative AI (GenAI) has become a business imperative, with organizations across industries racing to integrate this transformative technology. Leaders and experts have outlined various strategies for effective implementation (Chamorro-Premuzic, 2024), from identifying the right business problems to fostering a proactive mindset. Others advocate for a shift in perspective (Farri & Rosani, 2025)—moving beyond treating GenAI as a simple automation tool to leveraging it as a strategic thought partner.

Despite these insights, organizations continue to struggle with GenAI adoption. The transition is far from plug-and-play (Tse et al., 2024). Many companies find themselves grappling with fundamental structural mismatches, as existing organizational processes are not designed (Baier et al., 2024) to accommodate the iterative, dialogue-driven nature of GenAI.

A common response to this challenge is launching pilot projects, often driven by tech-driven innovation teams eager to test AI capabilities in controlled environments. Consider, for example, a B2B sales team participating in such an initiative. A salesperson already managing an overwhelming stack of digital tools (Tobias et al., 2023), may see a GenAI-driven pilot as just another tech experiment—one that demands time without offering immediate value. From their perspective, this isn’t about productivity; it’s about being asked to test a tool that doesn’t yet prove its worth. Their job hasn’t changed—they still need to sell. But now, they’re expected to divert attention to a system they don’t fully trust. And trust, as with any new technology, isn’t automatic and this slow adoption.

To ensure meaningful GenAI adoption, companies need a structured approach that aligns with modern and emerging management principles. One promising direction lies in co-creation and platform thinking (Trabucchi & Buganza, 2022 2023, 2025)—leveraging collaborative ecosystems to drive adoption and value creation. Let’s explore this through an inspiring, real, case.

Fujitsu’s Approach to GenAI Adoption

Initially, the Technology Vision and Thought Leadership team at Fujitsu designed four distinct GPT models, each catering to different internal audiences. This approach quickly revealed the limitations of a traditional method. Potential users distrusted the GPTs and viewed them as an additional burden to their jobs. This dynamic led Fujitsu to a bolder strategy based on co-creation as the key to adoption.

Rather than enforcing a top-down mandate, the company developed an internal initiative aimed at exploring the potential of large language models (LLMs) in knowledge sharing and business process optimization. At the heart of this effort was a personalized AI experience—each employee had the opportunity to create and customize their own GPT-based assistant, tailoring it to their specific role and needs. This was made possible through features available in modern GenAI tools like Copilot Studio or MyGPTs, which allow users to design AI assistants tailored to their specific needs. Rather than imposing AI adoption from the top down, Fujitsu embraced employee-driven innovation, leveraging co-creation as a mechanism to push adoption organically. Employees had the autonomy to develop tools that fit their workflows, ensuring higher engagement and long-term usability.

At the same time, Fujitsu leveraged Microsoft Viva Engage (previously Yammer) to create a space where employees could share their GPTs, test them, and receive valuable feedback to enhance their value and functionality.

In other words, Fujitsu shifted from a linear approach with a central unity developing GPTs to support departments in their processes to a marketplace (platform) where employees, rather than just being passive adopters of the new technology, could develop their own solutions and share them with whoever might benefit from them (Parker and Van Alstyne, 2005; Gawer and Cusumano, 2014; Teece et al. 2022).

This open, collaborative approach fostered a dynamic internal community of AI users. Employees became co-creators rather than passive adopters, actively shaping how the technology evolved within the organization. As a result, the initiative not only improved operational efficiency but also facilitated the widespread dissemination of a GenAI culture and strategic knowledge across teams. Fujitsu’s case illustrates how platform thinking—emphasizing co-creation, ecosystem engagement, and iterative refinement, even within the company—can drive more meaningful and scalable adoption of GenAI.

Technology Vision GPT: A Practical Case

One of the most impactful cases of these GPTs co-created and shared between organizational units was the “Technology Vision GPT”.

The Fujitsu Technology Vision and Thought Leadership team plays a critical role in shaping how the company’s innovations translate into business value. By developing new use cases for Fujitsu’s technology, this unit provides strategic insights that enable the sales team to connect innovation with customer needs. However, ensuring that sales representatives have access to the right use cases at the right time and with the right level of detail has always been a challenge, complicated by competing priorities, time constraints, varying levels of technical understanding, and differing customer expectations.

The Technology Vision GPT was built upon the Custom GPT Sharing Platform to bridge this gap. It is an AI-driven platform designed to aggregate, organize, and deliver use cases dynamically. Sales teams can now search, ask questions, and refine their understanding of new technological applications in real-time. This ensures that, before meeting a potential client, they can quickly identify the most relevant use case, validate key details, and confidently tailor their pitch.

But the impact goes both ways, working as an internal platform. While the GPT clearly serves the sales team, it also serves the Fujitsu Technology Vision and Thought Leadership team. By analyzing engagement data—such as which use cases receive the most queries, where clarifications are frequently needed, or which topics are underexplored—the team could gain valuable insights to refine their content, identify unmet needs, and improve their communication strategy.

In essence, the Technology Vision GPT operates as a two-sided platform, creating value for both the supply side (Fujitsu Technology & Service Vision team) and the demand side (sales teams). It enhances knowledge-sharing and fosters a continuous feedback loop, ensuring that the use cases being developed are not just available—but actionable, impactful, and aligned with business realities.

By designing the platform as a service for both sides (Muzellec et al., 2015; Trabucchi and Buganza, 2023), rather than a tool that required extra effort to adopt, Fujitsu successfully embedded AI into core business interactions. Technology Vision and Thought Leadership Teams were no longer providers in the knowledge management process but customers to be served. This shift allowed them to experience both the cost and the value of participation, strengthening their engagement and alignment with the broader business strategy.

A survey conducted among employees clearly shows the validity of this approach. A clear trend emerged: while many custom GPTs were designed to facilitate collaboration between different teams, the vast majority of GPTs actively used were those that supported individual work—even if originally created by someone else. This underscores a critical principle of platform design: when AI tools are built not just for automation, but for enabling collaboration and knowledge-sharing, and when all the sides involved perceive themselves as customers of the tool, adoption becomes a natural outcome rather than a forced initiative.

Key Principles for Adopting GenAI Through Platform Thinking

For GenAI to become a transformative force within organizations, adoption must be driven by meaningful interaction, not mere implementation. Fujitsu’s experience demonstrates that platform thinking provides a structured approach to embedding GenAI in a way that enhances collaboration, efficiency, and business impact. To replicate this success, organizations should embrace the following principles:

1. Enable, Don’t Impose

One of the most powerful aspects of GenAI is the ability to create custom, context-aware AI assistants tailored to specific needs. Rather than enforcing rigid, top-down AI applications in isolated pilot projects, companies should focus on empowering employees to create and adapt their own GPT models (Chesbrough, 2003). Partnering with technology providers that enable this flexibility—through tools like Copilot Studio or MyGPTs—allows users to shape AI tools around their workflows, rather than the other way around. This bottom-up approach not only increases adoption but also fosters innovation as employees discover unexpected but valuable use cases.

2. Customers, Not Suppliers

A core principle of platform thinking is that every participant in the ecosystem should have a strong reason to engage (Muzellec et al., 2015). Organizations should avoid treating AI adoption as a supplier-driven initiative, where employees are merely expected to consume predefined AI applications. Instead, they should design AI systems as services that provide clear, immediate value—whether by making someone’s job easier, facilitating knowledge sharing, or enabling employees to support each other’s work. When AI adoption is framed as a way to serve internal stakeholders rather than demand effort from them, engagement becomes a natural outcome.

3. Agile, Not Stage-Gate

Traditional IT rollouts follow a stage-gate approach, where solutions are planned, built, and deployed in structured phases. However, with radically advancing technologies like GenAI, experimentation is key. By enabling employees to create, test, and share their own AI assistants, organizations accelerate the discovery of meaningful applications. Hands-on engagement fosters rapid iteration, ensuring that the most valuable AI use cases emerge organically through practice, rather than being dictated by predefined plans (Magistretti & Trabucchi, 2024).

References

Baier, DeLallo & Sviokla. “Your Organization Isn’t Designed to Work with GenAI.” Harvard Business Review Online. (2024).

Chamorro-Premuzic. “7 Strategies to Get Your Employees On Board with GenAI.” Harvard Business Review Online, (2024).

Chesbrough, H. W., & Appleyard, M. M. “Open Innovation and Strategy”. California Management Review, 50/1 (2007): 57-76.

Farri & Rosani. “How AI Can Help Managers Think Through Problems.” Harvard Business Review Online. (2025).

Gawer & Cusumano. “Industry platforms and ecosystem innovation.” Journal of Product Innovation Management, 31/3 (2014): 417-433.

Magistretti & Trabucchi. “Agile-as-a-tool and agile-as-a-culture: a comprehensive review of agile approaches adopting contingency and configuration theories. Review of Managerial Science, 19/1 (2025): 223-253.

Muzellec, Ronteau, & Lambkin. “Two-sided Internet platforms: A business model lifecycle perspective.” Industrial marketing management, 45 (2015): 139-150.

Parker & Van Alstyne. “Two-sided network effects: A theory of information product design.” Management science, 51/10 (2005): 1494-1504.

Teece, Pundziene, Heaton, & Vadi. “Managing Multi-Sided Platforms: Platform Origins and Go-to-Market Strategy.” California Management Review, 64/4 (2022) 5-19.

Tobias, Riley, Giblin, Gregory-Hosler. “Sellers Are Overwhelmed by New Technology.” Harvard Business Review Online. (2023).

Trabucchi & Buganza. “Landlords with no lands: a systematic literature review on hybrid multi-sided platforms and platform thinking.” European Journal of Innovation Management, 25/6 (2022): 64-96.

Trabucchi & Buganza. “Platform Thinking: Read the past. Write the future.” Business Expert Press. (2023).

Trabucchi & Buganza. ”The Digital Phoenix Effect.” (2025).

Tse, Esposito, Goh & Lee. “Why Adopting GenAI Is So Difficult.” Harvard Business Review Online. (2024).

Statement on the usage of GenAI

ChatGPT was used to support the editing and refinement of the manuscript’s language and structure. All content, ideas, interpretations, and conclusions are entirely the work of the authors, who take full responsibility for the integrity and originality of the manuscript.