AI agents for Oracle Database users
Oracle Private Agent Factory · Oracle AI Database 26ai
With Oracle Private Agent Factory we build AI agents that answer from your documents, query your data and handle repetitive tasks. They run next to your Oracle Database and under your security rules, with a local language model or a model in the cloud.
Our principle
Your data.
Your model.
Your rules.
Data stays in your Oracle Database, and agents work within your access rules.
The challenge
Company knowledge is scattered across documents, databases and people's heads. Because finding it is slow, people turn to public AI tools — and the data leaves the company, with no rules and no trace.
01Time
The answer exists in a policy, a contract or a database. But half an hour passes before someone finds it.
02Privacy
Personal, financial and business data do not belong in public chatbots. Regulation and internal rules are clear about that.
03Control
Use outside approved tools: nobody knows who asks what, or where the data goes.
04Adoption
A prototype is quick to build. Everyday use needs security, access rights and integration with existing systems.
You don't need another chatbot. You need AI that works by your rules, with your data.
The solution
Oracle Private Agent Factory is Oracle's platform for building AI agents and part of Oracle AI Database 26ai. Agents run right next to your database, under the same access rules as your systems. SmartQ sets up the platform, connects it to your sources and builds your first agent with you.
Architecture
Your Oracle Database and documents
↓
Private Agent Factory, vector search, Select AI and a language model of your choice
↓
An agent in your business process
Capabilities
✓ Answers from your documents and cites the source
✓ Queries data in natural language
✓ Runs multi-step workflows
✓ Uses PL/SQL, REST and MCP tools
✓ Works inside your applications via API
Environments
→ On your own servers
→ In an isolated (air-gapped) environment
→ In your Oracle Cloud tenancy
→ On Oracle Database@Azure, AWS or Google Cloud
→ Wherever your Oracle Database runs
Privacy and control
For banks, the public sector, healthcare and industry this is often the first condition for adopting AI. Data stays in your Oracle Database, and you choose the language model to fit your requirements: entirely in-house or with a trusted cloud provider.
When regulation requires it, the language model and the embedding model run on your infrastructure. Questions, documents and answers never leave your environment.
An agent sees only what the user is allowed to see. Row- and column-level permissions, data masking and audit trails stay in force.
You can also use a model from a cloud provider, such as Azure OpenAI or OCI Generative AI. Data stays in your database; the model receives only the context needed to answer.
Use cases
The biggest gain is where people look for the same kind of information several times a day or repeat the same steps by hand.
01Knowledge
Policies, guidelines, contracts and procedures in one place. An answer in seconds, with a link to the source.
02Data
“What are sales by region this quarter?” The agent builds the query, table, chart and explanation — also for colleagues who don't know SQL.
03Finance
The agent compares invoices with contracts and rules, flags deviations and leaves an audit trail.
04Compliance
Answers only from approved documents. Every answer can be checked against its source.
05Support
HR or IT support and customer answers based on your content, not on general knowledge from the internet.
06Processes
The agent reads data, calls a PL/SQL procedure or a REST service and hands the result to the next step — or to a person for approval.
Business benefits
✓ you already keep your data in an Oracle Database
✓ regulation or internal rules keep your data out of public AI services
✓ you want local language models, or at least full control over the model
✓ your people often look for answers in documents and procedures
✓ you want to adopt AI quickly, but in an orderly and transparent way
Less searching and less manual work. More time for the work AI can't do.
Data and documents become usable for colleagues who don't write queries or know the database structure.
Give your people an approved environment. Access management, execution tracking and audit trails replace uncontrolled use of public tools.
Build on the Oracle Database you already have and know. Integrate agents into existing applications via REST — including APEX.
Private Agent Factory is part of Oracle AI Database 26ai and requires no additional licence. With the appropriate licences and active support with upgrade rights, the upgrade to 26ai is also included at no extra licence cost. You only pay for infrastructure and language models.
See the starter package ↗Starter package
We don't start with a long project. We set up the platform in your environment, choose a real use case together and deliver it on your data as a proof of concept. At the end you get an adoption blueprint with recommendations for production — and your team gets the know-how to build agents on its own.
SmartQ Oracle Private Agent Factory
EUR 18,900
✓ Private Agent Factory installed in your environment
✓ Data sources and language models connected
✓ Proof of concept on a real use case with your data
✓ 15 person-days of consulting and development
✓ “Next steps” workshop: adoption blueprint and recommendations for production
✓ Hands-on workshop for your team
Proof of concept
→ that data access is secure and under control
→ that answers are verifiable and backed by sources
→ how well the chosen model works with your data
→ what a single task costs
→ how the agent fits into a real process
We are an Oracle partner, and our consultants have on average 15 years of experience with data and analytics. We deliver Private Agent Factory, Select AI and MCP in practice and share what we learn as we go.
Intro call
In an intro call we pick the first task you'd like to simplify and check what your environment needs for it. No obligation.
Book an intro call ↗TELEPHONE
ADDRESS