Ackuity GenAI Security Gateway
Manage role based access, apply security and privacy to data exchange, control actions, get visibility and more in your Gen AI interactions- be it RAG, function call, AI Agents or text2 SQL pipelines.
Ackuity has three product modules as below
Secure Function Calling
Control how Text-to-SQL engines in GenAI applications interact with data warehouses and databases. Ackuity gives you a layer of security before these SQL queries are executed onto final systems.
Act as Tool
Ackuity can securely execute actions on enterprise systems using the function call parameters generated by GenAI applications.
Content Safeguard
Filter or transform the returned data of a function call for privacy, business rules, and industry regulations.
Authentication
Use Ackuity to centralize and enforce industry standards like OAuth and SAML authentication for all function or agent interactions.
Threat Management
Detect and respond to API security threats listed within the OWASP top 10.
Authorization
Limit data retrieval and other actions on enterprise systems based on roles and attributes of users and objects.
Visibility
See data on function calls, interactions between users, and applications over time.

Secure Agents & Tools
Control GenAI applications that perform Function Calling by using an agent-tool model. Apply the same security and privacy features as given above for only the agent model. In addition, control when a tool can be called by an agent based on user and query context.

.png?width=77&height=77&name=settings%20(1).png)
Agnostic to LLM models and supports GenAI frameworks, including Azure AI Studio, AWS Bedrock, GCP Vertex, LangChain, and LlamaIndex.
.png?width=77&height=77&name=settings%20(1).png)
Supports common enterprise & SaaS applications, including Salesforce, ServiceNow, Microsoft O365, Slack, HubSpot, Workday and more.
Secure Text2SQL
Control how Text-to-SQL engines in GenAI applications interact with data warehouses and databases. Ackuity gives you a layer of security before these SQL queries are executed onto final systems.
RBAC
Restrict SQL queries based on user roles.
ABAC
Restrict SQL queries or modify retrieved data based on user and data attributes.
Content Safeguard
Filter or transform the returned data including masking, tokenizing, and redacting fields for privacy, business rules, and industry regulations.
Threat Management
Detect and respond to injection and data leakage attacks.
Visibility
See queries, interactions between users, and databases over time.
.gif?width=1920&height=1080&name=Text2sqlGIF1-ezgif.com-loop-count%20(1).gif)
A technical variation of Chat2Database, where the inference pipeline for structured data involves API queries to data warehouses. Ackuity provides the same security and privacy features, while also enforcing robust authentication and detecting API threats.

.png?width=77&height=77&name=settings%20(1).png)
Agnostic to LLM models, and supports GenAI frameworks, including Azure AI Studio, AWS Bedrock, GCP Vertex, LangChain, and LlamaIndex.
.png?width=77&height=77&name=settings%20(1).png)
Supports common enterprise data platforms, including Azure Synapse, Amazon Redshift, Snowflake, Databricks, PostgreSQL, MySQL, and Oracle.
Secure RAG
RAG (with Vector Search)
Control GenAI applications where the RAG pipeline uses an intermediate search tool for hybrid semantic and vector queries. Apply the same security and privacy features of RAG with vector search.
RBAC
Restrict SQL queries based on user roles.
Threat Management
Detect and respond to data poisoning attacks.
ABAC
Restrict vector retrievals or modify retrieved data based on user and document attributes.
Content Safeguard
Filter or transform the returned vectors with masking, tokenizing, and redacting fields for privacy, business rules, and industry regulations.
Visibility
See data on vector retrievals, queries between, users and document repositories over time.

RAG (with Hybrid Search)
Control GenAI applications where the RAG pipeline uses an intermediate search tool for hybrid semantic and vector queries. Apply the same security and privacy features of RAG with vector search.
- Supports common search tools, including Azure AI Search, Amazon Kendra/Q, and Google Vertex AI Search.

.png?width=77&height=77&name=settings%20(1).png)
Agnostic to LLM models and supports GenAI frameworks, including Azure AI Studio, AWS Bedrock, GCP Vertex, LangChain, and LlamaIndex.
.png?width=77&height=77&name=settings%20(1).png)
Supports common document stores, including Microsoft SharePoint, OneDrive, Azure Files, Azure Blob Storage, AWS S3, AWS WorkDocs, and Google Drive.
.png?width=77&height=77&name=settings%20(1).png)
Supports common Vector Databases, including CosmosDB, Pinecone, Chroma, Weaviate, PostgreSQL, Amazon DynamoDB, and Redis.
How Ackuity works for you
Deploy
Launches as a SaaS application with a single policy enforcement module deployed as a container in your cloud or as a VM in your data center.
Configure
Includes out-of-the-box policies for security and privacy, and lets you write custom policies.
Analyze
Offers custom dashboards and on-demand reports for continuous monitoring and analysis.
Respond
Continuously enforces policies, monitors for violations, and allows for rapid one-click responses.