Summary:
1. AI agents are being deployed in production faster than expected, with companies seeing tangible results.
2. The trend is towards a multi-cloud, multi-model strategy, offering flexibility and optimization for cost and performance.
3. Enterprises are focused on solving real business challenges with AI, rather than chasing after artificial general intelligence.
Article:
The era of real, deployed agentic AI is here, and it’s already reshaping how businesses operate. Companies like Intuit, Capital One, LinkedIn, Stanford University, and Highmark Health are quietly putting AI agents into production, tackling concrete problems, and seeing tangible returns. The trend of deploying AI agents in customer-facing applications is accelerating at a breakneck pace, with 68% of enterprise companies already adopting agentic AI. This rapid shift is being fueled by tangible results, with companies like Intuit deploying AI agents to automate workflows and seeing businesses getting paid faster and in full.
The move towards a multi-model and multi-cloud strategy is evident, as enterprises seek the flexibility to choose the best tools for the job. IBM’s model gateway and Zoom’s three-tiered model approach reflect this desire for flexibility and optimization for cost and performance. Enterprises are getting smarter about how they use different models for different tasks, creating a powerful but constrained ecosystem that puts pressure on profitability.
While tech leaders discuss the dawn of superintelligence, enterprise practitioners are focused on solving real business challenges. Highmark Health is using LLMs for practical applications like multilingual communication, while Capital One is building teams of agents for tasks like risk evaluation and auditing. The travel industry is also adapting to new search paradigms enabled by LLMs, focusing on the customer experience rather than technology for its own sake.
The age of AI agents is transforming how teams are structured, with a shift towards small, nimble, and empowered teams. The future of AI teams lies in their ability to adapt to the rapidly evolving landscape of AI technologies and applications, focusing on practical solutions to real-world business problems. Summary:
1. Small teams of three to four engineers are considered most effective for rapid testing of product hypotheses.
2. Engineers are now becoming both builders and managers of AI agents, requiring skills in clear communication and strategic thinking.
3. The shift towards sandboxed development is fostering rapid innovation within a controlled environment to prove value quickly.
Rewritten Article:
In the fast-paced world of enterprise AI, the consensus is clear: small, agile teams of three to four engineers are the most effective for testing product hypotheses quickly. Varun Mohan, CEO of Windsurf, a leading agentic IDE, emphasized the benefits of this small team structure at a recent event. This approach allows for rapid experimentation and avoids the pitfalls of larger groups, ensuring efficient progress in AI development.
A significant shift is underway in the industry, where engineers are now expected to not only build AI agents but also manage them effectively. Companies like GitHub and Atlassian are recognizing the need for engineers to oversee fleets of agents, requiring a new set of skills that focus on clear communication and strategic thinking. This evolution is essential for guiding autonomous systems towards success in a rapidly changing landscape.
One key factor driving innovation in AI development is the growing acceptance of sandboxed development. Andrew Ng, a prominent figure in the AI space, advises leaving safety, governance, and observability to the end of the development cycle. While this may seem counterintuitive for some, the goal is to foster rapid innovation within a controlled environment to demonstrate value quickly. This approach aligns with the industry trend, as a recent survey found that 10% of organizations adopting AI have no dedicated AI safety team, highlighting a willingness to prioritize speed in the early stages of development.
Overall, these insights paint a clear picture of the evolving landscape of enterprise AI. Companies are transitioning from broad experimentation to focused, value-driven execution, deploying AI agents with a strategic mindset. The discussions at Transform 2025 showcased the industry’s readiness to embrace AI technologies, despite facing challenges and learning curves along the way. Early adoption and continuous learning are key to staying ahead in this rapidly evolving field.
For a deeper dive into these themes and further analysis from the event, consider listening to the full discussion on the recent podcast featuring independent AI developer Sam Witteveen. Additionally, the main-stage talks at VB Transform are available on YouTube, providing valuable insights into the future of AI in enterprise. Stay tuned for more coverage and updates from the event.
As a gesture of gratitude to our readers, we have opened early bird registration for VB Transform 2026 at just $200. This event is where AI ambition meets operational reality, offering a unique opportunity to engage with industry experts and stay ahead of the curve in the world of enterprise AI. Reserve your spot now to ensure you don’t miss out on this invaluable experience.
