Cohere North
by Cohere
Privately deployable enterprise AI workspace for agents, search and workflow automation — run it in your VPC, on-prem or air-gapped
Cohere North is an enterprise AI workspace that combines grounded search, document and content generation, custom agents and multi-step workflow automations in one platform. It is built for regulated organisations — banks, telcos, public sector — that want agentic AI but need to run it inside their own VPC, on-premises or air-gapped infrastructure rather than a shared SaaS cloud.
Cohere North is Cohere's all-in-one agentic AI platform for the enterprise, announced as generally available on August 6, 2025. It bundles three capabilities: Discover (search and retrieval grounded in the organisation's own data, from simple Q&A to decision support), Create (co-authoring documents, summaries, tables and charts), and Automate (custom agents built in Agent Studio and, since July 27, 2026, North Automations — plain-language, scheduled, multi-step workflows with loops, branching, per-step model selection, plan mode, version tracking, human approval checkpoints and token-consumption monitoring). North runs on Cohere's Command family of models plus its Embed and Rerank retrieval models, connects to workplace tools such as Gmail, Slack, Salesforce, Outlook and Linear, and supports Model Context Protocol (MCP) servers for industry-specific tools. Its core differentiator is deployment: it can run in a customer's private VPC, on-premises, in air-gapped environments, or on Cohere's isolated Model Vault inference service, and Cohere told TechCrunch it can run on as few as two GPUs. Cohere cites GDPR, SOC 2 and ISO 27001 compliance, granular access controls and agent autonomy policies. Named users include RBC (which runs a customised 'North for Banking'), Dell, LG, Ensemble Health Partners, Palantir and CoreWeave. Cohere, founded in Toronto in 2019, had raised about $1.6B at a $7B valuation by September 2025, agreed in 2026 to combine with Germany's Aleph Alpha, and was reported in September 2026 to be in talks to raise $2-3B at a $20B valuation. North competes with Microsoft Copilot, Google Gemini Enterprise, ChatGPT Enterprise and Glean, but targets buyers whose data cannot leave their perimeter.
CIOs and heads of AI at regulated enterprises (banking, telecom, healthcare, government) who must keep data and inference inside their own perimeter but still want a Copilot-class agent and search workspace.
A single agent, search and automation platform that can be deployed privately — VPC, on-prem or air-gapped — on modest GPU footprints.
At a Glance
- Category
- Enterprise Platform
- Pricing
- Contact for pricing
- Target Market
- CIOs, CTOs, Heads of AI, Enterprise IT, Regulated industries
- Deployment
- Hybrid, Self-hosted, Multi-cloud
- Founded
- 2019
- Headquarters
- Toronto, Canada
Key Features
- ✓Private deployment options
Runs in a customer VPC, on-premises, air-gapped, or on Cohere's isolated Model Vault, so sensitive data never leaves the perimeter.
- ✓North Automations
Plain-language, scheduled multi-step workflows with loops, branching, plan mode, versioning and approval checkpoints for governed agentic processes.
- ✓Per-step model selection
Lets builders choose a different model at each workflow step to balance cost against quality instead of overpaying on every call.
- ✓Agent Studio
Build customised AI agents that act across connected tools and data sources, tailored to a team's own processes.
- ✓Grounded search and retrieval
Uses Cohere Embed and Rerank models to answer questions from internal data, reducing hallucination versus ungrounded chat.
- ✓Connectors and MCP support
Built-in connectors to Gmail, Slack, Salesforce, Outlook and Linear plus MCP servers let agents reach existing systems without custom glue.
- ✓Governance and usage analytics
Granular access controls, agent autonomy policies, usage analytics and token monitoring give IT audit-ready visibility and cost control.
Capabilities
Use Cases
- •Sales call preparation
An automation consolidates Salesforce, Slack, Gmail and Notion data into presentation material before each prospect call, saving reps manual research.
- •Customer success morning briefings
A daily tracker pulls HR systems, schedules and Slack/Salesforce data to generate a morning briefing for customer-success managers.
- •Natural-language analytics
Marketing operations teams ask plain-language questions that North translates into SQL against BigQuery, delivering insights without an analyst queue.
- •Secure banking assistant
A bank deploys North privately, as RBC did with North for Banking, so employees use genAI on confidential client data inside the firewall.
- •Customer support transformation
Support teams ground answers in internal knowledge; Cohere cites CoreWeave transforming its customer support with North within 90 days.
Ideal For
Best For
- ✓Banks and financial institutions that need genAI assistants inside a private or air-gapped environment
- ✓Sovereign and public-sector deployments where data residency rules out shared SaaS AI
- ✓Enterprise-wide search and Q&A grounded in internal documents with access controls
- ✓Multi-step back-office automations (sales prep, CS briefings, analytics-to-SQL) with human approval checkpoints
- ✓Organisations standardising on Cohere Command, Embed and Rerank models for RAG
Not Ideal For
- ✗Small teams or startups — North has no self-serve price and enterprise licensing is reported to start at five figures a year with committed spend
- ✗Teams whose main workload is frontier-level coding or heavily multimodal generation, where independent reviews rate OpenAI, Anthropic and Google models ahead of Cohere's
- ✗Organisations already deeply invested in Microsoft 365 Copilot or Gemini Enterprise that do not need private deployment
Integrations
Deployment
Market Analysis
Pros
- ✓Genuinely private deployment (VPC, on-prem, air-gapped) that most Copilot-class rivals do not offer
- ✓Named regulated-industry references such as RBC, Dell, LG and Ensemble Health Partners
- ✓Strong retrieval stack (Embed, Rerank) for grounded enterprise search
- ✓Governance built into automations: approvals, versioning, token monitoring
Cons
- ✗No self-serve or published pricing for North; enterprise contracts reportedly start at five figures with committed spend
- ✗Independent reviews rate Cohere's models behind OpenAI, Anthropic and Google on raw performance, coding and multimodal work
- ✗No verified user reviews on G2 or Capterra yet, so buyers must rely on vendor references
- ✗Cohere faces a February 2025 copyright lawsuit from 14 publishers over training data
Pricing
North (Enterprise)
Contact for pricing
- ✓Agents, search, Create and Automations
- ✓Private VPC, on-prem or air-gapped deployment
- ✓Custom enterprise pricing via sales
Model Vault (dedicated inference)
From $2,500/mo
- ✓Isolated dedicated inference for Embed, Rerank and Parse models
- ✓Hourly rates from $4.00/hr
North has no published list price; Cohere's pricing page routes it to sales for custom enterprise quotes, and third-party analysis reports licensing starting at five figures annually with a committed minimum spend. Private deployment adds your own GPU costs, while Model Vault dedicated inference is published from $4.00/hr ($2,500/month).
Security & Compliance
Sources
This page was written from 9 sources, 5 on domains other than cohere.com.
- 1.cohere.com — northvendor
- 2.cohere.com — north gavendor
- 3.cohere.com — introducing north automations ai workflowsvendor
- 4.cohere.com — pricingvendor
- 5.techcrunch.com — coheres new ai agent platform north promises to keep enterpr
- 6.betakit.com — coheres valuation hits 7 billion usd following 100 million r
- 7.pymnts.com — ai startup cohere targets 20 billion dollar valuation fundin
- 8.en.wikipedia.org — Cohere
- 9.aimadetools.com — what is cohere north
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Pigment
AI-native integrated business planning with Modeler and Analyst agents, AI-generated planning apps and an MCP server
Red Hat AI
Run any model on any accelerator across hybrid cloud, with safety evidence and cost attribution built in
Genesys Cloud
Agentic orchestration for customer experience — AI agents, employees and systems on one governed platform
VMware Tanzu Platform
Production AI agents inside your own private cloud, with a deny-by-default runtime