By the time most companies figure out how to deploy one AI agent, Lua wants them to have an entire workforce of them.
The London-based startup announced Wednesday that it has closed a $5.8 million seed round led by Norrsken22, with participation from Flourish Ventures, Y Combinator, Phosphor Capital, and 20VC, among others.
The funding will go toward expanding Lua’s developer community and its global network of implementation partners.
The pitch is straightforward: stop treating AI agents like one-off experiments and start building them the way you’d build a team.
The Problem With How Companies Deploy AI Today
Most businesses, when they talk about deploying AI agents, are really talking about pilots — narrow, brittle tools that solve one problem, in one context, for one team. Scaling beyond that usually requires expensive engineering talent or vendor lock-in to platforms with opaque pricing models.
Lua’s founders, Lorcan O’Cathain and Stefan Kruger, saw this problem up close while scaling a fintech business in East Africa — O’Cathain as COO, Kruger as CTO.
Before starting Lua, O’Cathain ran the Africa business for Zephyr Management, a VC and private equity fund operating across Africa, India, and Sri Lanka. Kruger was VP of Engineering at Paystack before its acquisition by Stripe.
Their frustration wasn’t just with the tools. It was with the economics. “Most agent platforms compound this with black box tooling and per-outcome pricing,” O’Cathain said. “The more your agents succeed, the more you pay, with no pathway to improving your agent economics.”
Lua is built on the opposite model: teams own their agents, own their data, and — the theory goes — compound efficiency over time instead of just compounding costs.
One Platform, Two Kinds of Users
What Lua is building is, in its own framing, an opinionated full-stack agent platform — one that handles infrastructure, model orchestration, data pipelines, channel integrations, and monitoring. Businesses only have to supply the business logic and choose which integrations their agents need.
The platform is accessible through a CLI for developers who want granular control, and through a natural language interface for non-technical teams who don’t. Critically, both types of users are working on the same agents, in the same environment. A customer success manager and a software engineer can collaborate on the same agent workflow without translating between tools.
The company says a team can have a functioning agent workforce — coordinating handoffs between agents and humans, integrated into existing systems — within hours of getting started.
Since launching its developer platform in October 2025, Lua says revenue has grown nearly 30% week-over-week. In February 2026 alone, more agents were built on Lua than during the entire period from launch through January.
Why Norrsken22 Bet on Lua
Norrsken22, the Africa-focused venture fund that led the round, made the bet partly because of where Lua has already been deployed. The platform has seen adoption across Africa, Asia, the US, and Europe — giving the company a volume of real-world usage data and pricing intuition that’s genuinely hard to replicate from a standing start.
“The founders fundamentally understand how agent and human workforces need to collaborate to get work done,” said Lexi Novitske, General Partner at Norrsken22. The fund, which backs technology companies with operations in or meaningful ties to Africa, sees Lua’s global footprint as a structural advantage rather than a distraction.
Notable angel investors in the round include Henri Stern, CEO of Privy; Kaz Nejatian, CEO of Opendoor; and Med Benmansour, CEO of Nuitee.
Building the Workforce Layer for the Agent Era
The broader bet Lua is making is that the companies winning the next decade won’t just be the ones that use AI — they’ll be the ones that manage AI workforces with the same intentionality they bring to hiring, onboarding, and retaining people.
That framing positions Lua less as an AI tool and more as an HR system for non-human employees. It’s a bold metaphor, but it captures something real about where enterprise AI is headed.
The question isn’t whether businesses will run on agents. It’s whether they’ll own the infrastructure — or rent it, indefinitely, from someone else.
Learn more about other African tech startups on Labari Insights, our data repository for tech in Africa: insights.techlabari.com
