Fae — have we built a Lovable for Enterprise?

Tools like Claude Code and Codex are fantastic. Give them to a developer and productivity goes up immediately. They become, in their own box, hyperproductive. At least measured in code produced. But problems show up quickly when we try to scale it.

We at Wizardworks announced a few weeks back that we no longer charge for programming. That means we have to solve what the tools don't — because if every developer sits in their own silo, all we create is person-dependence and software written to personal taste, but at higher speed. That's a problem.

The tools, as mentioned, are built to make an individual developer hyperproductive. Not to give an organization control or to streamline software development. How do we handle the challenge of multiple-people-to-multiple-agents-to-multiple-projects? Who owns the context? Where's the traceability? How do we ensure AI-generated code meets compliance requirements? Who code-reviews when everyone is producing at full tilt? How do we focus all this hyperproductivity?

Those are the questions we've spent many, many (many!) cognitive cycles solving. And that's why we developed (and develop) Fae — Full Agentic Enterprise.

It's also a bit nerdily fun internally that we lean into the Wizardworks theme of magical creatures.

So what is Fae?

Fae is software, agents, processes, and structure for introducing agentic AI into an organization's development processes. We've built our own software so our agents can be as effective as possible in complex environments. All the tedious plumbing around how teams of agents and humans collaborate is in place.

In concrete terms, that means a combination of autonomous agents — triggered by various events — and interactive agents that collaborate with a team of humans in real time via Teams, Azure DevOps, and local environments via the terminal. At the same time, fully autonomous agents plan, implement, code-review, security-scan, and watch costs. In parallel. Round the clock if needed.

But it would be wrong to call Fae just software. It's scar tissue and know-how for setting up agentic flows in an organization. Guardrails, processes, adoption. The kind of thing that takes months (for humans, at least) to understand and build, and that can't be copied by downloading a tool.

We have customers who just want software delivered. And we have customers who want us to help them introduce agentic AI into their existing processes where their own developers are deeply involved in working with agentic AI. Fae is our vehicle that supports both.

Are human consultants still needed?

Both yes and no. Fae can run fully autonomously — requirements on one side, software on the other. But we advocate that there be a partner who takes responsibility for delivery and handles what still requires human-to-human interaction. AI-first with human accountability. There are things you really shouldn't delegate to an autonomous agent (looking at you, Entra ID).

We also have customers who absolutely could, but don't want to. A bit like how we ourselves are about accounting (receipts and VAT, ugh). Just like our customers, we'd rather have someone take responsibility.

Do I need to install anything?

No. Fae is a framework and platform on our side that handles agent orchestration, context, security, and so on. We adapt how Fae interacts with our customers based on their systems, processes, and requirements. Fae is, as mentioned, not just AI agents and software — it's also our know-how and structural capital for how (in our experience) to introduce agentic AI in an organization.

Why Wizardworks specifically?

Maybe a bit cheeky. But we think our long (broad?) experience of IT in larger organizations, where we've been involved at the leadership level, helps us. We also have experience driving (and selling) startups. That combination — enterprise complexity and startup mentality — is what made it possible to build something that packages agentic AI for organizations with real requirements.

How does this affect your existing business model?

So far it's the same. We've just taken agentic AI for Wizardworks to the next level — to be even more AI-first and hyperproductive in our deliveries. We use Fae for our customers. We even use Fae to build Fae. Which is completely, completely insane. The benefit for us and for our customers is that as Fae evolves and is built out, we can offer better and new capabilities.

In closing

Hand on heart, Lovable for Enterprise is probably the wrong analogy — but maybe it got your attention? What we've actually built (and are building) is a framework and a platform for autonomous agentic software development.

But that doesn't sound quite as cool.