Draft replies with citations
Every answer links back to the exact document section it came from. Citations include the document name and relevant passage.
Cortex - AI support copilot
Draft support replies from your trusted docs and tickets, with citations your team can verify.
Cortex is an AI support copilot, not a chatbot builder. Connect your knowledge base and Freshdesk, and Cortex retrieves the right sources, drafts a reply, and cites exactly where the answer came from - before your team ever sends it.
ticketWhat is the refund policy?
retrievepolicy.md section 4.2 + terms.md section 8
draftGenerated from 3 cited chunks
reviewawaiting agent approval before send
CitationsEvery draft cites its source
Review before sendA human approves every reply
Full audit logEvery generated answer is logged
EU-hostedHetzner infrastructure
The problem
Support agents search too much. AI chatbots hallucinate too easily.
Support agents spend real time searching documentation, old tickets, and internal procedures to answer questions customers have asked a hundred times before. General-purpose AI chatbots do not know your internal knowledge, and a confident wrong answer sent to a customer is worse than no answer at all.
Cortex solves this at the retrieval and drafting level. You ingest your knowledge base and connect Freshdesk. Cortex retrieves relevant passages, drafts a reply, and cites its sources. Your agent verifies and sends. Humans stay in control.
What Cortex is
An AI support copilot, not a chatbot.
What Cortex is
- A retrieval layer that finds relevant context for each ticket.
- A draft-and-cite step your team reviews before sending.
- An audit trail for queries, drafts, keys, and admin actions.
- An API endpoint your existing code can talk to.
What Cortex is not
- An autonomous ticket-resolution bot that sends without review.
- A general-purpose AI chat interface.
- A tool that sends your data to a shared model for training.
- A replacement for your support agents.
What Cortex can do
Core capabilities are built. Advanced interfaces are in internal beta.
Knowledge bases
Upload PDFs, Markdown, HTML, and plain text files. Connect external sources via webhooks or scheduled sync. Cortex chunks, embeds, and indexes everything automatically.
Human review before send
Cortex never sends a reply on its own. Every draft is queued for a person to check, edit, and approve first.
Hybrid retrieval
Cortex combines vector similarity, Postgres full-text search, and trigram similarity with configurable RRF weights. Optional cross-encoder re-ranking is available behind configuration.
Scoped API keys
Create keys scoped to read, chat, ingest, or full access. Rotate one integration without affecting the rest.
Audit log
Document uploads, assistant queries, API key usage, credential access, admin actions, and GDPR requests are logged per project.
OpenAI-compatible API
Point existing OpenAI client code at your Cortex project endpoint and get retrieval-augmented answers from your own knowledge base.
Fine-tuning (advanced, later-stage)
For teams with specialised domain language once the retrieval and drafting workflow is proven. Not required to start a pilot.
How it works
From ticket to verified answer, with a human in the loop.
Cortex never sends a reply on its own. It drafts. Your team reviews, verifies the sources, and sends.
Ticket arrives
A customer question comes in through Freshdesk (Zendesk next).
Cortex retrieves
Cortex searches your trusted docs, tickets, and procedures for relevant context.
Cortex drafts a reply
A draft answer is generated from the retrieved sources, not from memory.
Citations attached
The draft links back to the exact document and passage it came from.
Agent reviews
A support agent reads the draft and the sources before doing anything else.
Agent sends
The agent edits if needed, then sends the reply from Freshdesk.
Logged for audit
The ticket, sources, draft, and outcome are logged for review and improvement.
Prove it during the pilot
Designed to reduce search time and speed up first response.
We do not claim results before you have measured them on your own tickets. The pilot exists to track these metrics on your data, so you decide with evidence, not a demo.
Search time per ticket
Track how long agents spend searching docs and old tickets before Cortex, and after.
First response time
Measure how quickly a draft is ready for review once a ticket arrives.
Draft acceptance rate
Track how often agents send a Cortex draft with little or no edit.
Tickets with a useful draft
Measure the share of tickets where Cortex produces a usable starting point.
Edit / reject rate
Track how often agents rewrite or discard a draft, and why.
Citation accuracy
Spot-check whether cited sources actually support the drafted answer.
Escalation reduction
Track whether fewer tickets need senior or specialist escalation.
Agent onboarding time
Measure how quickly a new support agent becomes productive with Cortex drafting from day one.
The first packaged Cortex workflow
Symphony is the Freshdesk workflow powered by Cortex.
Symphony connects your Freshdesk workspace to Cortex. It reads incoming tickets, retrieves relevant context from your knowledge bases, and surfaces AI-drafted replies with citations your support team reviews before sending. It is not a separate product - it is Cortex, packaged for the support use case.
Example workflow
- Freshdesk webhook receives a new ticket.
- Cortex retrieves relevant documents and previous procedures.
- Suggested reply includes structured citations.
- Agent reviews, edits, and sends from Freshdesk.
- The interaction is logged for auditability and improvement.
Who uses Cortex
Built for teams that need answers, not guesses.
Support teams
Problem: Support agents search through documentation, old tickets, and procedures to answer repeated customer questions.
With Cortex: Cortex drafts answers from your knowledge base with citations the agent reviews before replying. This is the first Cortex workflow.
Product teams
Problem: Product knowledge is scattered across Notion, Confluence, Jira, repo docs, and changelogs.
With Cortex: Ingest the material and ask questions about features, decisions, and specs with source-backed answers.
Developer teams
Problem: You need AI in your product, but building a reliable RAG stack takes months.
With Cortex: Use the OpenAI-compatible endpoint with scoped API keys, streaming, webhooks, and no shared model training.
For developers
OpenAI-compatible. Drop-in ready.
If your application already uses OpenAI, you can point it at Cortex and keep the same client pattern. Retrieval, citations, and project isolation happen behind the endpoint.
from openai import OpenAI
client = OpenAI(
base_url="https://api.cortex.canopy.pm/v1",
api_key="crtx_sk_...",
)
response = client.chat.completions.create(
model="cortex-assistant",
messages=[{"role": "user", "content": "What is the refund policy?"}],
)
# Citations are included in the responseModel providers
Bring your own keys.
Cortex works with your existing OpenAI or Anthropic subscription. Your provider key is encrypted at rest with AES-GCM envelope encryption and is never returned in API responses or logs.
Supported now
- OpenAI
- Anthropic
Coming later
- Mistral
- Custom OpenAI-compatible endpoints
Data ownership
Your knowledge. Your project. Your data.
Every document, embedding, conversation, and API key in Cortex is scoped to your product instance. Your knowledge base data is never used to train shared or platform-level models.
Good to know
- SSO/SAML is planned for Q1 2027, not available today.
- Fully self-hosted deployment is not a standard option yet.
- Cortex is for static or periodically synced knowledge, not live database queries.
- Fine-tuning is optional and advanced - most pilots do not need it.
Cortex support pilot
A concrete, working pilot - not vague beta access.
We promise a working pilot within 14 days. Pricing is locked in for the full 90-day pilot. You leave with real numbers: search time, draft acceptance rate, citation accuracy, and more.
What's included
- SetupFrom €750
- MonthlyFrom €250/month
- Duration90 days
- IntegrationFreshdesk first
- DataDocs, old tickets, procedures, knowledge base content
- OutputReviewable AI draft replies with citations
FAQ
Questions technical buyers ask first.
Is Cortex a RAG system?
Yes. Cortex uses retrieval-augmented generation with hybrid retrieval: vector similarity, full-text search, and trigram matching before generation.
Does Cortex send replies automatically?
No. Cortex drafts a reply and attaches citations. A support agent reviews, edits if needed, and sends it. Cortex does not send on its own.
We use Freshdesk. How does the pilot integration work?
Symphony connects your Freshdesk workspace to Cortex over webhooks. Incoming tickets trigger retrieval against your knowledge base, and a draft reply with citations appears for agent review.
Do you support Zendesk?
Freshdesk is the first integration. Zendesk is planned next - contact us if that is a blocker for your pilot.
Do I need to fine-tune a model?
No. Most teams use the default retrieval and drafting pipeline. Fine-tuning is optional, advanced, and only useful for specialised domain vocabulary later on.
Can I use Cortex without an OpenAI subscription?
You need a model provider API key. Cortex supports OpenAI and Anthropic. You bring your own key and pay that provider directly.
Is Cortex GDPR-compliant?
Cortex provides GDPR erasure and export endpoints, EU infrastructure, project isolation, and no shared model training. A DPA is available on request.
What happens if I cancel?
Project data is retained for 30 days after cancellation. You can export your data before deletion.
Book a pilot
Ready to start? Tell us about your support workflow.
We prioritise teams with a Freshdesk workspace, a real backlog of docs and old tickets, and a willingness to work closely with us for the first 90 days.