AI-Driven Evolution: How Felmo Redefines Tech-Enabled Services
Explore how Felmo, a mobile vet service, leverages Generative AI and Agentic Engineering to boost efficiency and elevate performance across its tech teams.
Key Insights
-
Insight
Felmo, a mobile veterinary service, exemplifies a 'Tech-Enabled' business model where technology is not the product but a critical enabler for distributed operations. This approach optimizes efficiency in unique constraints, such as mobile documentation and route optimization for over 100 vets across Germany.
Impact
This model demonstrates how strategic technology integration can transform traditional service industries, leading to significant operational efficiencies and market differentiation for businesses with mobile workforces.
-
Insight
Felmo's tech team has rapidly adopted AI tools, with an estimated 50% of code pushes being AI-generated. This shifts developer focus from mundane coding to higher-level tasks like planning and specification, and has led to a significant increase in test coverage and overall development speed.
Impact
This shift suggests a future where developers act more as 'AI orchestrators,' accelerating development cycles, improving code quality, and potentially reducing overall engineering costs while increasing throughput.
-
Insight
The decision to switch primary AI coding tools from Cursor to Cloud Code was driven by budget constraints and the pursuit of more cost-efficient token usage from models like Anthropic's Opus 4.6. This pragmatic approach balances advanced features with economic viability.
Impact
Highlights the critical need for tech leaders to continuously evaluate AI tool costs versus performance to ensure sustainable AI adoption and maximize ROI, influencing procurement strategies for AI services in budget-conscious environments.
-
Insight
Felmo is actively reducing the 'Human in the Loop' in its development and QA processes, automating trivial code reviews and aspects of QA that were previously too complex or expensive for traditional automation. This allows human experts to focus on higher-value tasks.
Impact
Suggests a future where AI significantly streamlines testing and review processes, freeing human experts to concentrate on complex, high-impact issues, potentially leading to faster release cycles and higher quality software.
-
Insight
The increased efficiency from AI enables product developers to assume greater end-to-end responsibility, from initial specification and design collaboration to coding, deployment, and direct monitoring of business outcomes. This blurs traditional role boundaries within the organization.
Impact
This shift implies a move towards smaller, highly empowered, cross-functional teams capable of rapid iteration and direct impact on customer value, potentially making organizations more agile and responsive to market demands.
-
Insight
AI will significantly raise the average performance standards across various domains, creating a 'superhuman' effect where even junior professionals can operate at a higher baseline. This elevates competition across the board, including the quality of undesirable content like spam.
Impact
Foresees a future where baseline competence is elevated, requiring individuals and organizations to continuously innovate and differentiate themselves at much higher levels to remain competitive and relevant in an AI-augmented world.
Key Quotes
"Felmo ist ein mobiler Tierarzt. Wir sind deutschlandweit mit mehr als 100 Tierärzten aktiv. Und das wollte ich auch unbedingt sagen, wir sind kein Tech-Start-up, aber wir sind Tech-Enabled, die Tierärzte. Also das Modell funktioniert deswegen, weil wir Technologie haben."
"Ich würde behaupten, dass die Hälfte unseres Codes, also die, sagen wir mal, die Hälfte unserer Pushes AI-generiert sind. Und mindestens drei von unseren Kollegen, also vielleicht sagen wir mal, also sagen wir mal ein Drittel bis 40 Prozent der Kollegen machen hauptsächlich generierten Code und natürlich mit Anpassung."
"Der Human in the Loop wird weniger ein Thema. Ich würde tatsächlich sagen, was eigentlich passiert, ist, dass der Produktentwickler, also nicht der Frontend oder der Backend-Entwickler, sondern der Produktentwickler eben viel mehr Verantwortung auch übernehmen wird von der Spezifikation, also vielleicht auch in der Kollaboration mit dem Designer, bis hin tatsächlich zum Business Outcome."
Summary
The AI-Fueled Startup: Felmo's Journey into Generative Engineering
In an era defined by rapid technological shifts, startups are uniquely positioned to harness cutting-edge innovations. Felmo, a mobile veterinary service, exemplifies this by strategically integrating Generative AI and Agentic Engineering not just into its core tech, but throughout its entire tech-enabled operational model.
Felmo's Tech-Enabled Foundation
Felmo operates over 100 mobile veterinary teams across Germany. While not a "tech-startup" in the traditional sense, technology is its fundamental enabler. From optimizing vet routes to streamlining mobile documentation and billing – processes traditionally time-consuming for on-the-go professionals – Felmo's lean central tech team of under 20 people uses technology to make their distributed vet services efficient and scalable. This approach underscores how vital digital tools are to transform and empower traditional service industries.
Transforming the Developer Workflow with AI
The adoption of AI tools within Felmo's product, tech, and data teams has been swift and impactful. Tools like Cloud Code, Cursor, ChatGPT, and Gemini are now integral. It's estimated that roughly 50% of Felmo's code pushes are AI-generated, a figure that continues to rise, especially for test generation. This shift means developers spend less time on rote coding and more on high-level planning, architectural design, and problem specification.
Strategic tool selection, driven by a balance of features and cost-efficiency, is crucial. Felmo switched to Cloud Code from Cursor due to better token usage economics from Anthropic's models, while still leveraging Cursor's Bugbot for automated code review. This pragmatic approach ensures sustainable AI integration within startup budget constraints.
Beyond Coding: AI's Broader Organizational Impact
AI's influence at Felmo extends beyond pure software development. It's used for:
* Data Analysis: Empowering analysts to generate complex reports and insights more rapidly, shifting them towards more "coding-adjacent" roles. * Marketing: Generating content in Felmo's brand voice for social media, CRM, and even experimenting with video generation to adapt to employee changes. * Operational Planning: Assisting non-tech teams with tasks like survey design, strategic planning (e.g., practice renovations), and general knowledge acquisition.
This broad application demonstrates AI's potential to augment capabilities across various functions, making a lean central team highly effective in supporting a decentralized operation.
The Future of Work: Superhumans and Elevated Standards
The integration of AI is not without its challenges, including the constant pressure to adapt to new models and tools, and the mental stress of continuous context switching. However, Felmo is witnessing a profound transformation: a reduction in the "Human in the Loop" for routine tasks like QA and code reviews. This frees human experts to focus on complex, critical areas, leading to faster development cycles and improved quality.
Stefan Schubert-Peters, Felmo's CPTO, predicts a societal "level up" where AI significantly raises the average performance standards across industries. Developers, armed with these tools, become "Superhumans," capable of higher-level work and broader responsibilities, blurring traditional role boundaries and fostering tighter, more interdisciplinary teams. This ultimately means increased competition, but also unprecedented opportunities for innovation and progress.
Conclusion
Felmo's experience underscores a critical truth: Generative AI is reshaping the enterprise, enabling leaner teams to achieve more, accelerating product development, and empowering every employee. By embracing AI as a catalyst for efficiency and innovation, organizations can navigate this new landscape, focusing on customer value and strategic growth, and ultimately, making a collective "level up" a reality.
Action Items
Establish dedicated internal forums, such as an 'AI-Corner,' and provide individual AI budgets to encourage widespread experimentation and adoption of AI tools across all teams. Regularly showcase AI-driven successes to foster a culture of innovation.
Impact: Fosters organic AI adoption, empowers employees to discover practical applications for AI in their roles, and increases overall organizational efficiency and preparedness for future tech shifts.
Implement a continuous evaluation process for AI models and tools, comparing cost-performance ratios to optimize budget and efficiency. Be prepared to switch providers or adjust usage based on evolving market offerings and internal needs.
Impact: Ensures AI investments yield maximum return, prevents cost overruns in the rapidly changing AI landscape, and maintains a competitive edge by leveraging the most efficient available technologies.
Actively leverage LLMs like ChatGPT and Gemini for data analysis, learning new technical skills, generating complex dashboards, and understanding existing internal processes or codebases. Encourage this use across analytical and technical teams.
Impact: Significantly reduces the time and specialized expertise required for data-driven insights and onboarding, democratizing access to information and accelerating decision-making across the organization.
Prioritize AI-driven automation for non-core but critical functions such as marketing content generation, operational planning (e.g., facility renovations), and general knowledge support. Extend AI's benefits beyond core tech to augment capabilities in support functions.
Impact: Extends the benefits of AI beyond the tech department, streamlines cross-functional workflows, and allows lean central teams to support growing decentralized operations more effectively with fewer resources.
Rethink team structures to empower product developers with broader, end-to-end responsibilities, blurring traditional lines between frontend/backend, development/design, and business outcome ownership. Leverage AI to manage the increased scope and complexity.
Impact: Cultivates highly agile, integrated teams capable of faster delivery and direct alignment with business goals, potentially leading to more impactful product development and reduced communication overhead.
Mentioned Companies
Felmo
5.0Central to the discussion, showcased successful and strategic AI integration across its operations and development.
Cloud Code
4.0Praised for its efficiency and cost-effectiveness, becoming the primary AI coding tool after switching from Cursor due to budget.
ChatGPT
4.0Widely used for generalist tasks, data analysis, learning new skills (e.g., Postgres), and generating marketing content.
Anthropic
4.0Its models (e.g., Opus 4.6) are highly valued for performance and cost efficiency when accessed directly, influencing tool adoption decisions.
OpenAI
4.0The underlying technology for ChatGPT and GPTs, which are extensively used within Felmo's marketing team for content generation.
Cursor
3.0Initially favored for its strong features and transparency, but faced budget constraints, leading to a partial switch to Cloud Code. Still used for specific functions like the Bugbot.
Gemini
3.0Utilized for data analysis and general queries, with positive remarks on its "no bullshit" compliance in certain interactions.
GitHub Actions
3.0Mentioned as a critical part of the automated deployment and testing chain, which is being enhanced by AI-driven review processes.