Dojo fuels an AI-native future, delivers 680+ agents in two months with Gemini

*Featured in this video:* Robert Howes, Chief Technology Officer at Dojo *Executive summary:* Dojo, the leading payments provider in the UK, partnered with Google Cloud Consulting Services on its journey to become AI-native. Using Gemini as a front door to AI at work, Dojo launched a company-wide enablement program to give its teams the skills, tools, and confidence to make AI a part of their daily routines. By lowering the barrier to entry, every employee can now turn their unique expertise into custom AI solutions in minutes — driving the creation of over 680 agents in just two months. These capabilities aren’t only empowering frontline business teams but also helping engineers build sophisticated multimodal agents at scale, including a Chargeback Agent that reduced dispute processing times by 70%, and a document parsing agent that cut onboarding and compliance workflows from hours to minutes. Gemini’s built-in governance and security safeguards, including tools like Agent Gateway, Model Armor, and Agent Identity, ensure Dojo can provide the same accountability, oversight, and boundaries as any member of its workforce. *Challenge:* In the fast-moving payments industry, speed and security are just table stakes, but agility is what shapes the future. On its mission to become AI-native, Dojo set out to lower the barrier to entry for AI across its entire organization, empowering every employee to help transform the way they work. *Solution:* After building its core platform on Google Cloud — including Google Kubernetes Engine, Cloud Spanner, BigQuery, and Looker — Dojo partnered with Google Cloud Consulting Services to give staff the skills and confidence to use AI everyday, including an education plan, deployment roadmap, technical support, joint governance and collaboration, and thought leadership. Using Gemini as its front door to AI at work, along with self-paced and instructor-led Google Skills training, Dojo enabled employees to turn everyday business expertise into custom solutions in minutes. Already, non-technical teams have delivered new AI tools, ranging from daily sales agents that save reps seven hours weekly to recruitment agents that shave off an hour per week from talent acquisition processes by standardizing job postings. In addition, engineers now have the tools they need to build multimodal agents at scale, such as the Chargeback Agent that cut dispute processing times on complex cases by 70% and a document parsing agent that accelerated onboarding and compliance workflows from hours to minutes. To govern this growing ecosystem, Dojo relies on Agent Gateway, Model Armor, and Agent Identity to maintain oversight, protect sensitive data, and enforce security and compliance policies. *Results:* With Gemini, Dojo has unlocked company-wide creation to deliver over 680 AI agents in just two months — the vast majority built by frontline business teams using natural language. By addressing key AI enablement and governance requirements early on, Dojo is paving the way to broaden its agentic ecosystem, with plans to explore advanced code generation tools and expand its use of Gemini into customer-facing experiences. *Key takeaways and highlights from our interview with Robert Howes, Chief Technology Officer at Dojo:* → “We're using the Gemini platform to build agents that are both no-code and high-code in nature. The ability to connect all of Dojo's key information sources means we have an accessible agent building capability that's also supported by Dojo-wide context.” → “When you can use natural language, it means anyone from HR to sales can turn their unique business experience into a custom AI agent in minutes. In two months, we created over 600 agents. The majority were built by our frontline business teams, right inside Gemini.” → “Agent Gateway acts as a critical layer between our core applications and Gemini. When combined with Google's Model Armor, it allows us to inspect all traffic — scrubbing PII, applying security policies on every single LLM call. Each agent workflow is bound using Agent Identity to a specific persona and set of appropriate permissions, all tracked in one place against Agent Registry.” → "It's really clear that Google is thinking about not just building agents out, but also enabling Dojo to trust running those at enterprise scale." *Google Cloud products used:* Gemini, including platform features such as Agent Gateway, Model Armor, and Agent Identity; as well as Google Kubernetes Engine (GKE), Spanner, BigQuery, and Looker. *Learn more:* → New Way Now: Dojo is a payments powerhouse, delivering 60% faster transactions with Spanner https://www.youtube.com/watch?v=rvcaA31GKHg