A new venture aiming to connect AI services across platforms could transform everyday commerce by letting virtual shopping agents compare prices, schedule deliveries and handle customer service across multiple retailers and websites.
The vision of /dev/agents, which just landed $56 million in funding at a $500 million valuation, points to a future where artificial intelligence (AI) helpers could independently manage tasks from restocking office supplies to scheduling home repairs — provided the startup succeeds in its ambitious goal of creating an operating system that lets different companies’ AI agents work together like apps on a smartphone. While tech giants like Microsoft and OpenAI are already rolling out specialized AI agents for tasks like healthcare scheduling and marketing, the lack of a common framework has limited their ability to coordinate across services.
“If the system is open and allows seamless integration of tools, it could drive down costs by reducing duplication of effort and making it easier for businesses to adopt AI,” Kevin Baragona, founder of DeepAI, told PYMNTS. “For example, instead of paying for multiple disconnected AI services, businesses could manage everything under one unified platform. But if it becomes a closed ecosystem controlled by a few players, they might charge premium prices, locking smaller businesses out. It really hinges on whether it’s built to serve the whole industry or just the creators’ bottom line.”
The startup, led by former Google VP David Singleton and other Android veterans, wants to create a foundation to let AI agents talk to each other. Right now, AI assistants from different companies work in isolation — imagine if apps on your phone couldn’t share information or work together.
The operating system /dev/agents plans to build would work like Android does for mobile apps, providing the basic rules and tools that let different AI services coordinate their actions. For instance, if you’re shopping online, one company’s AI assistant could smoothly connect with others, handling shipping, customer service and payments across different stores. The hefty $56 million in startup funding suggests investors see this as a crucial piece of infrastructure for AI’s future.
Interest in AI agents is booming. OpenAI revealed “Operator,” an AI agent launching in January to tackle complex tasks. Microsoft’s Ignite 2024 showcased agents for automating customer returns and shipping. Anthropic debuted the Model Context Protocol, linking AI assistants to live data. Samsung is integrating ChatGPT into Galaxy devices, enhancing AI capabilities.
Andrew Brooks, who runs the AI platform Contextual.io, told PYMNTS that AI agents are crossing a major threshold — moving beyond giving recommendations to making real-world decisions. His examples paint the picture: Instead of flagging angry customers for human review, AI agents can now read the situation across emails and chats and issue refunds directly. In warehouses, they don’t just predict inventory shortages — they place the orders themselves. The shift marks a fundamental change from AI that merely suggests solutions to systems that can independently execute complex business tasks.
“A specific live example for our clients is an AI Agent that can automatically approve pricing based on a prediction of wage rates in a given location,” he said. “This spans local job data and a Salesforce quotation process. Because these AI Agents are automating actions that would otherwise at least require a human, the impact is a reduction in staff required to perform the same decision-making process. Note that it is less about the action itself (approving a quote is a simple step) but rather about advanced decision-making incorporating multiple data sources that are being eliminated.”
As AI takes over more complex decisions, the debate shifts to who will shape how these tools work together. According to Andrey Meshcheryakov of consulting firm Recombinators, it will likely come down to one key question: Will major tech companies dominate the development of AI operating systems, or will open-source initiatives take the lead? He suggests that Big Tech firms could set the standards and use their vast resources and existing ecosystems to lock users into proprietary systems — much like how OpenAI has focused on consumer products.
“However, open systems, like the microagent projects, could push for decentralized, collaborative frameworks,” he said. “The balance will hinge on market demand for openness versus convenience and the willingness of smaller players to rally around open standards.”
We’re always on the lookout for opportunities to partner with innovators and disruptors.
Learn More