The Mark Cuban Podcast
Episode: North Cohere’s AI Agents: Disruption in Action
Date: August 17, 2025
Host: Mark Cuban (A)
Episode Overview
In this episode, Mark Cuban dives deep into Cohere’s latest enterprise AI agent platform, “North,” exploring its focus on security, enterprise deployment, and how it sets itself apart in an increasingly saturated AI agent market. Cuban analyzes Cohere’s strategy, the technical specs of North, its customizability, and its positioning against dominant players like OpenAI. He also provides critical insights into industry challenges, product competitiveness, and the importance of enterprise-specific solutions.
Key Discussion Points & Insights
1. Cohere's Evolving Position in AI
- Cuban acknowledges that Cohere initially lagged in LLM competitiveness but is now "pushing themselves back to the front" with the North agent platform.
- Notable Canadian branding—Cuban jokes about their naming:
"Their new AI agent platform is called North. And, you know, of course they're. They're up north, so I think that's some pretty good Canadian branding for you." (02:10)
2. What Are AI Agents—And Where Do They Stand?
- Discussion on how AI "agents" are conceptualized: not as full autonomous employees, but as tools to automate specific business tasks.
"There's this concept with agents that at some point we're gonna be like, hey, you're a marketing agent... make my sales go crazy. ...I think less and less this is where people are, and more and more it's... giving it tasks to accomplish." (05:20)
- Distinction made between agents performing single tasks vs. full workflow automation.
3. Security and Data Privacy as Core Differentiators
- Cuban breaks down enterprise concerns around data privacy and the prominence of security leaks in AI deployments.
- Cohere’s “North” promises secure, private deployments—can be run on-premises, within a company’s own infrastructure, or in air-gapped environments, rather than on Cohere’s cloud.
"This isn't a deployment where you're accessing their API. This is going to be on your own servers, your own firewalls, protected." (08:50)
- Emphasis on granular access control, agent autonomy policies, continuous red teaming, and independent security testing to meet international compliance standards (GDPR, SOC2, ISO 27001).
4. Enterprise Adoption and High-Profile Clients
- North is already piloted by RBC, Dell, LG, Ensemble Health Partners, and Palantir, showcasing trust among significant enterprise clients.
- Cuban notes:
"They said that they already piloted north with a bunch of customers. So RBC, Dell, LG... and also Palantir. So there’s some big players that are using this." (13:30)
5. Platform Features: Beyond "Just" Chatbots
- North’s primary features include chat and search, allowing users to handle queries, summarize meetings, write marketing copy, and search both internal and web data—standard for enterprise agents but executed with an emphasis on security and verifiability.
- It provides citations and a "reasoning chain of thought," so employees can audit how answers are produced:
"It will show the steps that it went through to get an answer. So if you're saying, like, hey, which customers in our database had the most problematic transactions, it would give you a list and then you can go look through how it actually came to that conclusion." (15:00)
6. Integration and Asset Creation
- Seamless integration with existing workplace tools like Gmail, Slack, Salesforce, Outlook, and custom in-house applications via Model Context Protocol (MCP).
- Advanced asset creation:
"North has a bunch of assets, a bunch of asset creation. It can make tables, it can make documents, it can make slideshows, a bunch of market research..." (19:00)
- Acquisition of Auto Grid allows Cohere to enhance market research and automation tools for enterprises.
7. Cohere’s Strategic Positioning: Tools over Models
- Cuban’s critical analysis:
"I don't think this is actually a super competitive product as a standalone product... they're not competing head on with OpenAI... They're basically trying to say... we have a whole bunch of really great tools for enterprise to use." (23:50)
- The real value lies in tailored enterprise tools wrapped around their own not-top-tier LLM.
- Raises a provocative industry question:
"It almost calls into question, unless they're about to train and release like some new insane LLM, they're not number one. And so it calls into question... what if any company went and took that open model, ran it themselves and built all the tools Cohere has and competed with them without having to train their own model?" (25:10)
8. Quotes – Cohere’s CEO and Co-Founder Nick Frost
- On LLM deployment and customer data:
"LLMs are only as good as the data they have access to. If we want LLMs to be as useful as possible, they have to have access to useful data and that means they need to be deployed in the customer’s environment." (09:55)
- On deployment versatility:
"We can deploy literally on a GPU in a closet that they might have somewhere." (11:15)
- On model augmentation vs. automation:
"As you build confidence by chatting to the model, there's like a smooth transition that happens between using this as an augmentation to using it as an automation." (21:50)
Memorable Moments
- Cuban’s playful appreciation of Canadian branding throughout the product names—North and Compass:
"It's so funny. I love this. Compass. North. This is great. This is a very, very Canadian." (18:00)
- Direct, critical analysis of the competitiveness of Cohere’s model and the broader implications for enterprise AI vendors in a rapidly commoditizing market.
Notable Timestamps
- 03:00 – Introduction to Cohere’s new "North" AI agent platform
- 05:20 – What AI agents can (and cannot) do today
- 08:50 – Private deployment and security focus explained
- 09:55 – Nick Frost’s insight on LLM deployment
- 13:30 – List of early major enterprise partners
- 15:00 – Reasoning chain of thought and auditability
- 18:00 – Multimodal capabilities and Canadian naming humor
- 19:00 – Asset creation and market research integration
- 21:50 – Building confidence: from augmentation to automation (Nick Frost quote)
- 23:50 – Cuban’s critical analysis of product competitiveness and market strategy
- 25:10 – Future threats from open models and Cohere’s needed investments
Final Takeaways
- Cohere's North is a compelling enterprise AI agent platform with a unique focus on security, private deployment, auditability, and integration—but it isn't positioned to beat the likes of OpenAI in raw LLM power.
- The company’s success seems rooted in building flexible, needs-focused enterprise tools atop a serviceable but not market-leading model—demonstrating that in enterprise AI, tools and integration can rival pure model dominance.
- Mark Cuban closes with a reminder: even if you’re not the best at everything, you can still reach multi-billion-dollar valuations by solving real user problems with the right partnerships and strategic focus.
