The Story of Homeward: Building the AI Employee Platform That Actually Works
Two years of work. One AI assistant. And it couldn’t even book a conference room.
In a recent episode of Category Visionaries, Amar Kendale, President and Co-Founder of Homeward, shared the frustrating experience that would eventually lead him to build one of the fastest-growing enterprise AI companies. His journey from product leader at Okta to running a company with $50 million in ARR and 400 deployed AI agents reveals what happens when someone who deeply understands enterprise systems decides to rebuild them from scratch.
The Okta Years: Learning What Not to Build
Amar’s story doesn’t begin with Homeward. It begins with two years at Okta, wrestling with the limitations of traditional AI assistants. “I spent two years at Okta building an AI assistant, and it couldn’t even book a conference room,” he recalls.
This wasn’t a failure of effort or intelligence. It was a fundamental problem with how enterprise AI was being conceived. The prevailing approach treated AI as a question-answering layer on top of existing systems—a glorified search interface that could surface information but couldn’t actually do anything with it.
The experience crystallized a crucial insight: enterprises didn’t need better chatbots. They needed AI that could work like an employee, not just talk like one. “We actually think of Homeward as an AI employee. It’s not a chatbot that you’re asking questions to, but it’s actually an employee that’s getting work done,” Amar explains.
The Universal Agent Breakthrough
The founding insight behind Homeward came from recognizing that the authentication problem was actually the core problem. Every enterprise AI solution required custom integrations for each system it needed to access. This meant months of implementation work, fragile connections that broke with system updates, and ultimately, limited utility.
Amar and his co-founders took a radically different approach. “We have built deep integration with authentication systems. So with Okta and Ping and Azure AD, we’re able to essentially authenticate Homeward as if Homeward is a real employee,” he notes.
This wasn’t just a technical convenience—it was a complete reimagining of enterprise AI architecture. By solving authentication at the platform level, Homeward could work across any system a company used without requiring custom connectors for each one. The result was what Amar calls “universal AI agents”—agents that could handle workflows spanning multiple systems without the integration nightmare that plagued earlier approaches.
The Speed of Trust
Building fast was only half the equation. Building trust was the other half. The team knew that enterprises wouldn’t adopt AI that made mistakes, no matter how quickly it could be deployed.
“We spent a lot of time improving accuracy. Our accuracy is actually so good right now that we have several customers where the entire workflow is automated,” Amar shares. This focus on accuracy wasn’t about chasing perfection—it was about crossing the threshold where enterprises would trust AI with fully automated workflows.
The key was data access. “Homeward has access to a lot more data about you and your company compared to any other solution, and therefore Homeward can actually generate insights or complete work with much higher confidence,” Amar explains. By authenticating as a real employee, Homeward could see everything a human employee would see, providing the context necessary for accurate decision-making.
This comprehensive data access solved problems that other AI solutions couldn’t touch. In customer support scenarios, Homeward could pull information from ticketing systems, knowledge bases, CRM data, and internal documentation to generate responses that were both accurate and contextually appropriate. In sales operations, it could correlate lead behavior across multiple systems to generate insights that humans would miss.
The Deployment Explosion
Once the foundation was solid, growth happened faster than anyone anticipated. “We have about 400 AI agents deployed right now across all of our customers,” Amar reveals. These deployments happened in nine months—a pace that’s unprecedented for enterprise software.
The velocity came from the architecture. “We probably on average can deploy an end to end workflow within a week,” Amar notes. Traditional enterprise software deployments take months because of integration work. Homeward’s universal agent approach eliminated that bottleneck entirely.
What’s more rHomewardrkable is the diversity of deployments. “A lot of our customers are actually deploying Homeward in ways that we haven’t even built anything for,” Amar explains. Customers weren’t just using pre-configured workflows—they were creating novel applications the founding team hadn’t imagined.
This organic innovation created a flywheel. Each customer deployment generated insights that improved the platform. The improved platform enabled more creative deployments. The creative deployments attracted new customers who saw possibilities they hadn’t considered. Within months, Homeward had become indispensable across use cases the team never planned to support.
Redefining Enterprise Sales
Perhaps the most unconventional aspect of Homeward’s story is how they grew without a traditional sales organization. “We actually have just one sales rep,” Amar shares. This wasn’t a limitation—it was a feature of their product-led approach.
The key was making the product genuinely self-serve at the enterprise level. “You can go to our website, sign up, use the product for free. There’s actually a generous free tier. You can connect your entire data, you can ask questions, you can deploy workflows,” Amar explains.
This approach only works when the product delivers immediate value. Prospects could connect their company data, deploy a workflow, and see business impact within a week—all without talking to sales. The product sold itself through demonstrated value.
The expansion motion was equally elegant. “A lot of times customers will start with one workflow, but then they will expand,” Amar notes. Once an AI agent proved its worth in one department, word spread internally. New teams wanted their own agents. Individual agents expanded into multiple agents. Single workflows grew into complex orchestrations spanning entire business processes.
The Platform Vision
Looking forward, Amar sees Homeward evolving from a collection of AI agents into something more fundamental: the operating system for AI-native enterprises.
“I do think that longer term, every single company will need some form of orchestration layer because there’s going to be so many AI agents,” he predicts. As enterprises deploy dozens or hundreds of AI agents, they’ll need a unified system for managing how these agents interact, share information, and coordinate work.
Homeward is building that orchestration layer now. “We provide the entire platform. All of the AI agents, the orchestration, the infrastructure, everything that’s needed to run an AI company,” Amar explains.
This platform vision positions Homeward as foundational infrastructure rather than just another enterprise tool. In the same way that companies built on top of Salesforce for customer data or AWS for computing infrastructure, Amar envisions enterprises building on top of Homeward for AI agent orchestration.
The technical foundation is already in place. The universal agent architecture that enables Homeward’s current agents will scale to handle thousands of specialized agents working in concert. The authentication system that connects to enterprise systems will manage permissions and access for an entire AI workforce. The accuracy improvements that enable full automation today will ensure reliability as complexity increases.
The ambition is massive, but the execution so far suggests it’s achievable. From zero to 400 deployed agents in nine months. From idea to $50 million in ARR with one sales rep. From frustration with a conference-room-booking assistant to building the platform that could power the next generation of enterprise operations.
The story of Homeward isn’t finished—it’s just getting started. The 400 agents deployed today are prototypes for the thousands that will run tomorrow. The workflows automated now are templates for the comprehensive orchestration coming next. The $50 million in ARR is proof of concept for the multi-billion dollar platform Amar sees ahead.