AI Agent Builder Gumloop Lands $50M From Benchmark

$50 million. That’s what Benchmark just bet on a two-year-old startup building AI tools for people who can’t code.

Gumloop closed a Series B Tuesday morning. The ai agent builder secured financing from Benchmark partner Everett Randle, who led the round. Nexus VP, First Round Capital, Y Combinator, Box Group, The Cannon Project, and Shopify participated.

Company wasn’t even raising.

“We decided this was the year to step on the gas,” co-founder Max Brodeur-Urbas told reporters.

The round values Gumloop higher than previous rounds, though exact valuation wasn’t disclosed. For Randle—who joined Benchmark last October from Kleiner Perkins—this marks his first deal at the iconic firm. Benchmark backed eBay, Uber, Dropbox. Now Gumloop.

Here’s what the company does: it lets non-technical employees build AI agents that automate work. No engineers required. Workers at Shopify, Ramp, Gusto, Samsara, Instacart, and Opendoor deploy agents through the platform. These agents handle multi-step tasks autonomously.

Employees share agents internally. That creates compounding effects.

“They get addicted, they start building more agents, and then all of a sudden, the whole company is AI native,” Brodeur-Urbas explained.

When I ran TaskFlow, we spent six months building internal automation tools. Burned engineering time we didn’t have. Gumloop’s pitch: skip the dev work entirely. Let every employee become a builder.

That’s the theory. Question is whether it works at scale.

Randle ran due diligence and found something interesting. One customer adopted Gumloop organically—no top-down mandate. The CTO gave employees full access to Gumloop plus two competitors. Six months later? Staff used Gumloop daily or weekly. The other tools sat untouched.

Why?

Minimal learning curve. “You can go in and start making agents and workflow automations immediately,” Randle said.

Most automation platforms require training, technical knowledge, or dedicated admins. Gumloop’s ai agent builder removes those barriers. That’s the edge Randle sees versus Zapier, n8n, Dust, and even Anthropic’s Claude Co-Work.

Model-agnostic approach matters too. Gumloop doesn’t lock users into OpenAI or Anthropic or Google. Teams choose whichever model performs best for specific tasks. As models evolve, flexibility wins.

Cost angle: enterprises hold credits across OpenAI, Gemini, Anthropic. They want to burn those credits efficiently. Gumloop lets them.

Brodeur-Urbas founded Gumloop mid-2023. Original vision: help workers automate repetitive tasks using AI. Back then, AI agents were experimental and error-prone. Technology matured. Product evolved.

Early plan? “Build a 10-person, billion-dollar company.” That changed. Enterprise demand surged. Now the company’s scaling sales and engineering teams aggressively.

Raising $10M isn’t success. It’s a liability. But Brodeur-Urbas is using this capital to build revenue infrastructure—sales teams that close enterprise contracts. That’s execution.

Benchmark partnership sealed the decision. Brodeur-Urbas called it a “no-brainer.” Benchmark’s brand opens enterprise doors. Their portfolio companies become potential customers and distribution channels.

Competitive landscape is brutal. Zapier owns workflow automation for SMBs. N8n serves technical users. Dust targets knowledge workers. Anthropic’s building native agent tools inside Claude.

Every AI lab will eventually offer agent-building features. Why wouldn’t they? The functionality sits one layer above foundation models. OpenAI, Google, Anthropic—all have incentive to own the stack.

Randle disagrees that threatens Gumloop. “Model independence is precisely what will keep attracting customers,” he argued. As models evolve, one outperforms another for specific use cases. Gumloop provides flexibility to switch.

Most VCs say that. Execution determines whether it’s true.

Here’s what matters: enterprise adoption velocity. If Gumloop spreads organically inside companies—like that CTO’s org where employees chose it over competitors—the ai agent builder platform becomes sticky. Network effects kick in. Shared agents create internal momentum.

If it requires top-down mandates and training programs? Zapier already won that game.

Randle views this as the biggest category in enterprise AI. “Enterprise automation is a massive pot of gold,” he said. He’s betting $50M that Gumloop captures meaningful share.

I’ve seen this movie before. Workflow automation tools promise to democratize technical work. Most fail because non-technical users hit complexity walls. Gumloop’s bet: AI removes those walls entirely.

Brodeur-Urbas claims employees get “addicted” to building agents. That’s the metric that matters. Not seats sold. Not MAUs. Addicted users who build daily.

Bootstrapped businesses don’t get headlines. They get profitable. Gumloop took VC money, so profitability trades for growth. Fair trade if enterprise customers pay and stay. Death sentence if they churn after pilots.

Typical SaaS companies target 120% net retention. Enterprise deals lock in annual contracts. Gumloop needs to prove agents deliver ROI that justifies renewals. Automation tools often promise savings but require maintenance overhead that kills value.

The company’s model-agnostic positioning helps long-term. Foundation models will commoditize. Whoever builds the abstraction layer that lets enterprises use any model wins infrastructure value.

Think about cloud: AWS, Google Cloud, Azure compete on compute. Terraform and Kubernetes became valuable by abstracting across providers. Same dynamic could play out in AI agents.

Randle joining Benchmark from Kleiner Perkins brought deal flow and conviction. This Series B represents his thesis: every worker needs AI superpowers. Gumloop’s no-code ai agent builder unlocks that potential.

For now, capital buys time to prove enterprise traction. Revenue growth, customer retention, and expansion revenue determine whether this becomes a sustainable business or another overfunded automation tool.

Most startups pivot 2-3 times before finding product-market fit. Gumloop already pivoted from basic task automation to full agent-building platform. That evolution matches AI capabilities maturing from 2023 to now.

Next milestones: enterprise sales team delivery, customer case studies showing ROI, and retention metrics proving agents stick. Benchmark doesn’t invest $50M for lifestyle businesses. They want unicorns.

Execution beats ideas. Every time. Brodeur-Urbas has capital and distribution. Now he executes.