Hawk AI™ · Natural Language to SQL
Ask your data in plain English — get governed SQL
Hawk AI turns a natural-language question into SQL, runs it, and charts the result — in seconds. And because it executes through DBHawk, every AI query obeys the same access control, masking, and audit as your team.
Runs in your infrastructure Every prompt audited Ask in any language 
The idea
Self-service analytics — without the security trade-off
Most AI-query tools ship your schema and data to a third-party API and hand back SQL that runs with unrestricted access. Hawk AI is different: it runs inside your own infrastructure and executes through DBHawk's policy layer, so natural-language querying stays fast for users and safe for security teams.
Question → SQL → chart
Ask in plain language. Hawk AI writes the SQL, executes it, and visualizes results as charts and tables instantly.
Governed execution
Generated queries run under the same object & column access control and dynamic masking as any user — no bypass.
Any language, any database
Phrase questions in the language you think in, across Oracle, SQL Server, Snowflake, PostgreSQL and more.
How Hawk AI runs a question
Fast for users, accountable for security
1
Ask in plain language
“Show refunds over $1k last month by store.” No SQL knowledge required.
2
Hawk AI writes SQL
Your chosen model — local via Ollama or Azure OpenAI — generates the query, in your environment.
3
DBHawk enforces policy
The query runs through access control and masking; the user only ever sees permitted, redacted data.
4
Chart & audit
Results render as charts and tables; the prompt, SQL and result are logged for compliance.
Why security teams approve it
AI that plays by your rules
Same access controls — AI queries respect object, column, and row policy.
Masking applies — sensitive fields stay redacted in AI results.
Full audit — every prompt, generated query, and result is logged to the real user.
Runs in your infrastructure — data and prompts don’t leave your environment.
Enforced through SSO — AI access is tied to the user’s identity and role.
FAQ
Natural language to SQL — questions
Does Hawk AI send my data or schema to a public AI service?
No. Hawk AI runs inside your own infrastructure. You bring your own model — a local model via Ollama or Azure OpenAI in your tenant — so prompts, schema, and results stay within your environment. There is no shared, multi-tenant API in the path.
Can AI-generated queries see data a user isn't allowed to?
No. Hawk AI executes through DBHawk, so its queries are subject to the same object- and column-level access control and dynamic masking as any user. The AI cannot return data the requesting user isn't permitted to see, and masked fields stay redacted.
Which AI models are supported?
You choose the model you trust: a self-hosted local model through Ollama, or Azure OpenAI. This "bring your own model" approach keeps you in control of cost, residency, and vendor choice.
Is AI activity auditable?
Yes. Every prompt, the SQL it generated, and the result are logged against the real user identity and secured through your enterprise SSO — the same audit trail you rely on for human activity.
Let your team ask questions — safely
See Hawk AI turn plain-language questions into governed, audited SQL and charts, running entirely in your environment.