TerraAmp KnowledgeAI Logo
SOC2-Aligned Architecture

Enterprise RAG Intelligence Layer.

Bridge the gap between raw data and actionable knowledge. Multi-tenant RAG-as-a-Service designed for high-compliance enterprise environments.

Schedule discovery callJoin Waitlist
KnowledgeAI Chatbot Interface

Used by teams in healthcare, legal, and cybersecurity

Built on pgvector and Claude — battle-tested in TerraAmp's own vCISO platform.

What customers get

A governed AI layer over the knowledge your teams already use.

KnowledgeAI turns documents, policies, procedures, tickets, and reference material into a cited assistant that can be deployed for employees, customers, and operations teams.

Faster answers

Reduce manual search across PDFs, folders, portals, and internal notes.

Cited trust

Give users the answer and the source trail needed to verify it.

Lower support load

Deflect repeated questions with a reusable assistant trained on approved content.

Operational visibility

See what people ask, where content is missing, and which workflows need improvement.

Built for teams that can't afford a hallucinated answer.

Every layer of KnowledgeAI — retrieval, governance, deployment — is designed for the compliance bar of regulated enterprises.

Cited retrieval

Every answer links back to the exact source paragraph, page, or row — so trust is verifiable, not assumed.

Multi-tenant isolation

Logical and physical separation per workspace. Vector indexes, encryption keys and audit logs are tenant-scoped.

Connectors that respect ACLs

SharePoint, Google Drive, Confluence, S3, Notion — permissions are mirrored, never flattened.

Hybrid retrieval engine

Dense + sparse + re-ranking, tuned per workspace. X% improvement in top-3 recall on a 500-document corpus vs. cosine-only search — methodology available on request.

Governance & audit

Immutable query logs, PII redaction, role-based prompts. SOC2-aligned architecture and GDPR ready.

Bring your own model

OpenAI, Anthropic, Mistral, or self-hosted via vLLM. Swap models without rebuilding indexes.

Solution depth

More than a chatbot: ingestion, retrieval, governance, and rollout in one operating model.

Knowledge preparation

  • Document upload and connector-ready architecture
  • Chunking and metadata strategy
  • Workspace and tenant scoping

Retrieval quality

  • Vector search for semantic matches
  • Keyword search for exact terminology
  • Re-ranking for better answer context

Assistant experience

  • Website and workspace chatbot
  • Grounded responses with citations
  • Conversation history for review

Access control

  • Role-based admin paths
  • Tenant-aware data model
  • Source permission strategy for enterprise connectors

Provisioning workflows

  • Admin-triggered tenant setup
  • API-driven tenant setup with full automation

Pilot analytics

  • Question patterns and usage review
  • Content gap identification
  • Production-readiness recommendations

Workflow diagram

How information becomes a trusted answer

The platform separates source content, retrieval, generation, and governance so the assistant can answer with confidence and show where the answer came from.

1

Ingest approved knowledge

Upload documents or connect approved repositories. Each file is normalized, chunked, tagged, and stored with tenant context.

2

Retrieve the right evidence

Hybrid search finds the most relevant snippets using semantic meaning, exact terms, metadata, and re-ranking.

3

Generate with citations

The assistant answers only from retrieved evidence and returns source references for verification.

4

Review and improve

Analytics show common questions, weak answers, and missing content so the knowledge base gets better over time.

SharePoint
Google Drive
PDFs / SOPs
secure ingestion
Chunk
Embed
Tag
hybrid retrieval

KnowledgeAI retrieval layer

tenant index + metadata + citations

Vector match
Keyword match
Re-rank
grounded response

Answer with source citations

delivered to web chat, teams, support, or internal portal

Response summary based on approved sources.
Sources: Policy.pdf page 4, SOP.docx section 2.1

Deployment blueprint

Start with a pilot, then scale to the right operating model.

KnowledgeAI supports a practical path from one assistant and one knowledge base to multi-department and enterprise-controlled deployments.

01

Pilot

Validate one use case, up to 500 documents, and a 30-day usage window.

02

Starter

Launch one assistant for a focused team or customer-facing workflow.

03

Business / Professional

Add departments, repositories, APIs, workflows, and analytics.

04

Enterprise

Move to dedicated environments, private vector stores, SSO, audit, and compliance controls.

Plans & Pricing

Flexible options tailored for pilots and full-scale production deployments.

Option 1: Pilot Engagement (Recommended)

AI Knowledge Assistant Pilot

$1,500USD FIXED FEE

Designed for fast procurement approvals.

Watch what the 30-day pilot delivers

Discovery, document ingestion, chatbot deployment, live pilot usage, and analytics review.

Discovery workshop
Knowledge ingestion
Up to 500 documents
Chatbot deployment
30-day pilot
Analytics review
Schedule Pilot Setup

Production Pricing

Starter

$299/month
Includes
  • 1 AI Assistant
  • 1 knowledge base
  • Up to 5 users
  • Up to 500 documents
  • Website chatbot
  • Source citations
Schedule a call
Target

Small clinics, Accounting firms, Law offices, Consultants

Best Seller

Business

$799/month
Includes
  • Multiple departments
  • Up to 25 users
  • Workflow automation
  • API access
  • Teams integration
  • Analytics
Schedule a call
Target

Financial services, EdTech, SaaS companies, Professional services firms

Professional

$1,499/month
Includes
  • Multi-department deployment
  • Advanced RAG
  • Multiple knowledge repositories
  • Custom workflows
  • Priority support
Schedule a call
Target

Healthcare groups, MSPs, Cybersecurity firms, Manufacturing

Enterprise

$2,999+/month
Includes
  • Dedicated environment
  • Private vector database
  • SSO & Audit logs
  • Compliance controls
  • On-premises options
Schedule a call
Target

Large enterprises & strict compliance environments

Questions, answered.

Still curious? Contact our team.

Is KnowledgeAI SOC2 compliant?
KnowledgeAI is built on a SOC2 Type II-ready architecture — immutable audit logs, role-based access controls, PII redaction, and tenant-scoped encryption keys are in place by design. Our formal SOC2 Type II audit is currently underway. In the meantime, we can provide a security architecture overview, our controls mapping against the SOC2 Trust Services Criteria, and a summary of our most recent penetration test on request. Contact us at security@terraamp.in or use the form on our Contact page.
How does KnowledgeAI handle GDPR?
KnowledgeAI's architecture is designed for GDPR readiness. Data is processed and stored within tenant-scoped environments with encryption at rest and in transit. We support data residency requirements, respond to data subject access requests (DSARs), and maintain audit logs that support your Article 30 records-of-processing obligations. A standard Data Processing Agreement (DPA) is available for all Business, Professional, and Enterprise customers — request it during your discovery call or via our Contact page. We are not a GDPR legal advisor; please review our Privacy Policy and involve your DPO for final compliance assessment.
Do you support HIPAA for healthcare deployments?
Healthcare teams on Professional and Enterprise plans can request a Business Associate Agreement (BAA). Our architecture supports the technical safeguards required under HIPAA — access controls, audit controls, transmission security, and automatic logoff — but HIPAA compliance is a shared responsibility. We strongly recommend a joint review with your compliance team before deploying with PHI. Contact us to begin that conversation.
Can we deploy in our own VPC?
Yes. Enterprise plan customers can deploy KnowledgeAI within their own Virtual Private Cloud (AWS, GCP, or Azure) or on-premises. This includes a private vector database, no outbound data egress to shared infrastructure, and the ability to bring your own LLM endpoint (including air-gapped or self-hosted models via vLLM). VPC deployment scoping is part of the Enterprise onboarding process — schedule a call to discuss your environment.
How does KnowledgeAI keep tenant data isolated?
Each workspace operates with strict logical and physical separation. Vector indexes, encryption keys, and audit logs are scoped per tenant — your data is never co-mingled with another organization's data, and tenant context is enforced at the retrieval layer, not just the application layer. Workspace administrators have full visibility into their own audit trail. Isolation architecture documentation is available for security review on request.
Which LLM providers do you support?
We support OpenAI (GPT-4o and o-series), Anthropic (Claude 3.5 and Claude 3 Haiku), Mistral, and self-hosted models via vLLM or compatible OpenAI-spec endpoints. Models are swappable without rebuilding your vector indexes. For air-gapped or regulated environments, we support fully self-hosted inference — no data leaves your perimeter. Model selection is configurable per workspace, so different departments can use different models within the same deployment.