The challenge
A large percentage of our client’s technology team is engaged in the software development life cycle (SDLC). Recognizing the potential of Generative AI early on, they moved to centralize AI capabilities by developing their own GenAI platform. However, initial results fell short of expectations. The platform focused on generic use cases and still showed potential in context understanding, process traceability and customization for practical use by product teams. Despite significant investment, adoption among senior developers remained low, as the platform did not deliver sufficient practical value in high-stakes environments. The rollout was further constrained by technical limitations, inconsistent data quality, and limited GenAI expertise.
The challenge was therefore twofold: to address justified skepticism among experienced IT professionals and to ensure that AI tools delivered measurable, day-to-day utility. The client wanted to move beyond experimentation and embed AI into the daily workflows of thousands of developers.
The approach
Instead of a rigid, top-down rollout, we implemented a GenAI coaching model alongside ongoing tool development as part of the overall strategy. Our engagement focused on equipping teams and key roles with the skills and confidence to apply AI meaningfully in their daily work, while at the same time translating practitioner feedback into clear improvement priorities for AI-enabled tooling. Our methodology was built around four pillars designed to embed AI into daily workflows:
- Enablement network: We supported the further development of existing enablement structures by strengthening peer-to-peer learning formats and expanding hands-on training concepts. This helped create distributed expertise within the organization and enabled teams to solve AI-related challenges locally while sharing learnings across units.
- Practitioner-driven improvement: By systematically collecting feedback from users, we helped prioritize enhancements that increased the practical usefulness of AI-supported tools. This ensured that further development focused on real workflow needs rather than isolated experiments. As a result the platform’s context engineering capabilities improved, ensuring that the AI tools had access to the right data sources without being overwhelmed by noise. The introduction of MCP enabled users to reach agentic workflows from their AI apps and to connect to external systems.
- Tool-agnostic guidance: Recognizing that a single platform cannot solve every problem, we provided transparent advice on when to use the internal AI platform versus specialized 3rd party tools. This built the necessary trust with the client’s technical teams. We drove collaboration with power users to identify AI use cases and focus technical development on the most important value drivers.
- Agentic AI implementation: We moved beyond static LLM prompts to agentic self-service. This allowed teams to orchestrate multiple agents for their workflows. By structuring complex tasks into coordinated steps, teams were able to reduce manual effort in documentation, analysis and knowledge sharing, while retaining human oversight where required.
"The platform grew to over a thousand users, supporting them with dedicated working sessions, community townhalls and further knowledge-sharing formats."
The result
The initiative contributed to a broader and more consistent AI adoption across technology teams. Tens of high-impact features were defined and implemented, supported by a structured development pipeline to sustain ongoing innovation. Within six months, we supported knowledge transfer to a growing group of AI Ambassadors, contributing to the enablement effort across the organization. The platform grew to over a thousand users as the overall program supported them with dedicated working sessions, community townhalls and further knowledge-sharing formats. Several cost-saving use cases were successfully embedded into daily team workflows.
Overall, the organization progressed from isolated experimentation towards more decentralized and sustainable AI usage. Teams were empowered to adapt their own processes, and AI evolved from an occasional productivity aid into a more integrated component of daily work. Beyond technical outcomes, the bottom-up enablement approach supported a cultural shift towards more transparent, collaborative, and responsible use of AI, as well as reduced the risk of shadow AI.
The AI Office transitioned from a centralized platform provider to a facilitator of shared learning and best practices. This shift strengthened stakeholder trust and accelerated sustainable adoption across the organization.
About the client
Our client is the international provider of IT services, responsible for the entire IT infrastructure and digital transformation of a European insurance company. This initiative represents their dedicated commitment to integrating AI across their software development and IT operations.