How Athena Intelligence unlocked audit firms and regulated enterprises with self-hostable workflow infrastructure

About

Athena Intelligence builds an AI agent platform for enterprise automation. The company targets highly regulated industries—audit firms, financial services, healthcare—where compliance requirements mandate private cloud or on-premise deployment. Without deployment flexibility, entire market segments would be inaccessible.

The problem

Brendon Geils, founder of Athena Intelligence, spent four years evaluating workflow orchestration platforms. The blocker was always deployment: could it run in a customer's Azure environment? In an on-premise data center? Embedded into another product?

"I had tested every product on the market. There was not very many products that were willing to be deployed on private clouds, self-hosted, on-prem, embedded."

— Brendon Geils, Founder

Most workflow companies built SaaS products where multi-tenancy was baked into their business model. Self-hosting was an afterthought. But for Athena, this blocked access to their entire target market. Major audit firms won't route sensitive financial data through third-party SaaS. Healthcare companies face HIPAA requirements. Financial services have data sovereignty policies.

"We would never go with a vendor that wouldn't make that possible in one, two, three years time because it's like putting all your eggs in a basket and then at scale you can actually never deploy that basket into the facilities that we want."

— Brendon Geils

The workflow engine was foundational—getting it wrong meant either abandoning regulated markets or facing expensive re-architecture later. Hatchet was too unstable. Airflow was too low-level without a UI. Most platforms treated self-hosting as an afterthought with poor documentation.

The solution

Windmill supported cloud, self-hosted, on-premise, and embedded deployments as core capabilities, not afterthoughts. The open-source foundation meant Athena could verify deployment feasibility before committing—critical for enterprise IT reviews.

"Everything worked out of the box. It was easy for me to learn. The technical people on our team took to it really quickly."

— Brendon Geils

For a platform company, reliability was existential. When infrastructure embedded in a customer-facing product fails, it's Athena's reputation on the line.

"When I wake up at night and I'm like, 'Oh, something's on fire'—it's usually, it's not, it's never a windmill problem. Windmill is probably some of the most stable parts of our business."

— Brendon Geils

The documentation covered edge cases other platforms hadn't anticipated—network segmentation, identity provider integration, resource constraints. This reduced the deployment friction that kills enterprise deals.

How it works

When Athena's AI agent encounters a task it can't handle, customers write a custom script. The agent helps write the code, the script is saved in Windmill, and it becomes a new tool the agent invokes automatically.

"Most of our usage is the script capability. We have an agent that uses a bunch of tools and capabilities. Whenever we don't have a tool, we recommend our customers go and build a script. The agent can help them write the code and then the agent uses that script as a new tool."

— Brendon Geils

For enterprise customers, Windmill runs embedded within Athena's platform in their own infrastructure. An audit firm deploying to Azure sees Athena's AI interface—Windmill is invisible infrastructure underneath, executing scripts within their security perimeter.

# Example: Audit firm writes this compliance check once

# Script runs in their Azure environment with their data

def check_audit_compliance(company_id: str, audit_year: int):
    records = fetch_financial_records(company_id, audit_year)
    compliance_score = calculate_compliance(records)
    return {
        "score": compliance_score, 
        "status": "compliant" if compliance_score > 0.95 else "review_needed"
    }

# AI agent invokes this automatically - data never leaves customer infrastructure

Athena's customers write custom Python scripts that run securely within their own infrastructure.

The result

Athena is now deploying to a major audit firm in Azure—the type of regulated enterprise customer that requires private cloud infrastructure and would be inaccessible with a SaaS-only platform.

"We have a very large audit firm that we're deploying to now in Azure. A lot of that is on top of Windmill."

— Brendon Geils

After nearly a year in production, Windmill serves dozens of customer organizations running 24/7 automation. The infrastructure has proven rock-solid.

"Windmill ranks among the most stable components of our infrastructure for performance, cost, scaling, uptime."

— Brendon Geils

Deployment flexibility unlocked access to regulated markets:

• Audit firms requiring private cloud with data residency guarantees
• Financial services with on-premise mandates for sensitive data processing
• Healthcare organizations needing HIPAA-compliant infrastructure
• Government contractors with air-gapped network requirements

Each market has substantial budgets and long-term contracts but is only accessible to vendors whose infrastructure can deploy into their environments.

Why this matters for platform companies

For companies serving regulated industries, deployment flexibility determines market access. Athena spent years searching for a workflow platform that could unlock these markets. Not for more features, but for architectural flexibility that could scale across deployment environments while maintaining production reliability.

The lesson for platform companies and ISVs: choose foundational technology based on the markets you want to access three years from now, not just the customers you can sign today.

Conclusion

Athena Intelligence's story reveals a simple truth: for platform companies targeting regulated industries, deployment flexibility isn't a feature—it's market access. SaaS-only tools lock you out of audit firms, financial services, and healthcare. Self-hosting as an afterthought creates deployment friction that kills enterprise deals.

Windmill's open-source foundation, production stability, and enterprise-ready documentation solved this. After four years evaluating workflow platforms, Athena found infrastructure they could confidently build on and deploy wherever customers required. Today, they're winning enterprise deals their competitors can't even pursue.