Anything’s 15-person team works alongside more than 150 AI agents to serve over 1 million users. Today, the system behind those agents is available to every company to build its own AI workforce.
Anything, the AI app builder used by over 1 million people, launched Skydive, a new platform that lets anyone at a company create and deploy AI agents for specific jobs.
Catch more HRTech Insights: HRTech Interview with Emma Lavelle, Chief Operating Officer, UneeQ
You don’t need permission from IT to hire a coworker. You just need to know the job, and then you can build the coworker who does it.
Skydive agents aren’t chatbots or workflows. They’re coworkers. Each has a name, a role, and its own cloud computer, so they can use all tools a team uses, a browser, a file system, and a terminal. They work across Slack, email, iMessage, desktop, and the web, learning how a team operates and carrying work from request to completion, even when their laptops are closed.
Anything built Skydive to run its own company. Its 15-person team now works alongside more than 150 agents across customer support, engineering, marketing, growth, sales, operations, and the office itself.
Now, any company can build the coworkers they’re missing.
Built for the work companies never get to
Emma handles customer support in Intercom. Anything’s support lead built her in a weekend after an acquisition tripled ticket volume overnight. He also built Ellis, who guides customers through App Store and Google Play submissions. Ellis helped 90 customers publish their apps in his first week. When Emma and Ellis need to coordinate, they talk in a group chat.
Otis runs engineering project management. He watches ideas surface in Slack and maintains the roadmap in Linear. Anyone on the team can DM him or tag him in a thread.
Grace investigates bugs using production AWS CloudWatch logs and Anything’s codebase in GitHub. Engineering created her for the rest of the company to use. With Grace, a support lead who had never written code now ships more pull requests than some engineers on the team.
Polly generates brand creative with OpenAI and Figma for Meta paid ads campaigns and manages their spend. Rice is now the company’s superhuman SDR, researching customers, prioritizing leads, drafting messages, and scheduling sales demos over email.
Canary monitors production around the clock. At 2 a.m. one morning, she alerted a co-founder to an outage, identified the root cause, and explained how to fix it. She was right.
Justin orders the office snacks with Instacart. Employees post what they want in Slack, and Justin places orders online. Nobody on the human team touches the process.
This is how Anything operates today: when a new need surfaces, the person closest to it can create an agent to take ownership of the work.
Trained by the people who understand the job
Most agent platforms centralize internal agent development with an engineering or automation team. It takes so much time that they often end up with 1 agent for the entire company. Skydive works differently.
The person closest to the work can now build the agent.
A support leader can create an engineering lead who fixes every support request. A growth lead can create an outbound team. An operations manager can create a coworker for the repetitive job that stole their focus from what’s most important.
Teams can begin with templates for roles such as chief of staff, customer support, SEO, or bug investigation, then teach each agent how their company actually works. Agents improve as they learn the team’s tools, preferences, and processes.
“Our mission has always been to make it extremely easy for regular people to build,” said Dhruv Amin, co-founder of Anything. “Apps were the first version of that. Agents are next. Inside Anything, they haven’t replaced a single hire. They’ve taken on jobs we never had the headcount for and given every person here leverage that used to require an entire team.”
From idea to agent without reinventing infra
Each agent runs on its own cloud computer with a browser, file system, and terminal. Skydive handles model access, sandboxes, permissions, security, and infrastructure, so teams can focus on defining the job rather than assembling the system underneath it.
Skydive supports more than 170 frontier and open-weight models. Administrators can manage access, data, usage, and spending centrally, while employees remain free to build agents for the work they understand best.
“The biggest surprise for me has been the power of letting people make their own agents rather than centralizing it,” Amin said. “When you give that power to people, I’m shocked at how quickly they spin useful agents up. By this time next year, I predict most of the job market will have the experience of building their own agents to take on work.”
Read More on Hrtech : Agentic HR: Can AI Become a Workforce Strategist Instead of Just an Automation Tool?
[To share your insights with us, please write to psen@itechseries.com ]