One of the most common blockers for businesses trying to deploy AI agents is not the model and not the budget. It is the data. Years of customer records, transactions, and operational history live inside SQL databases built long before anyone was thinking about AI. Moving that data or rebuilding around it is expensive, risky, and slow. RavenDB is betting that businesses should not have to.

On September 8, 2026, RavenDB launched Quill, a context layer that sits on top of existing SQL databases and makes them immediately accessible to production AI agents. No migration required.

What Quill Actually Does

Quill connects to an organisation’s existing PostgreSQL, SQL Server, or MySQL database and creates a live AI context layer on top of it. The layer stays synchronised with the source database using change data capture, so agents are always working with current operational data.

The key design decision is that Quill sits between the AI and the source system. The AI workload runs outside the SQL database entirely, so nothing extra touches production. The organisation controls exactly what the agent can and cannot see, independent of the source database’s own permissions. RavenDB describes this as governance built in by default rather than bolted on later.

The package includes search, retrieval, and the agents themselves. It is model-agnostic, meaning teams can use any AI model, switch providers, or run entirely on their own hardware.

The target deployment timeline RavenDB is promoting is weeks, compared to what the company describes as the 18 to 24 months typical of building a custom AI stack in-house.

The Problem This Solves

Most medium and large organisations are in the same position. Their most valuable data, the kind that would make AI agents genuinely useful, sits in SQL databases that predate the current wave of AI tooling by years or decades.

The standard advice has been to modernise the data layer first. That means migrating to a vector database, building retrieval pipelines, standing up an embedding service, and wiring it all together before a single agent can run in production. For a company with multiple business units and years of operational data, that roadmap is measured in quarters, not sprints.

Quill is designed to compress that timeline by leaving the system of record exactly where it is and adding the AI capability on top rather than underneath.

What This Means for Business

For business owners and operations leaders, Quill is relevant for a specific reason: it removes one of the main excuses for not starting with AI agents.

The “our data is not ready” conversation is common. It is often true, but the version of “ready” being imagined is usually a complete data platform overhaul. Quill represents a different category of solution, one that treats your current database as an asset rather than a liability.

If your business runs on SQL Server, PostgreSQL, or MySQL and you have been told that deploying an AI agent on your own data requires months of infrastructure work first, that assumption is worth revisiting.

The governance layer is also worth noting for organisations operating in regulated industries. Because Quill sits between the model and the data, access controls are defined at the Quill layer rather than relying on the database’s own permissions or trusting the model to respect boundaries it cannot actually enforce.

This is the kind of practical AI infrastructure story that tends to get less attention than model launches but has more direct impact on how quickly real deployments happen.

RavenDB has not published pricing. More database integrations beyond the initial three are described as forthcoming.

---

About Enterprise DNA: We help business owners understand and deploy AI that actually works. Our Omni Ops service builds AI agent workforces tailored to your operations, working with whatever data infrastructure you already have.