Loading page…
The partners will combine Cohere’s AI systems with PwC’s implementation and risk expertise, aiming to scale automation while preserving data control and human judgment.
Listen to this story
Cohere and PwC have structured their alliance around taking enterprise AI from pilots into governed workflows, with Canada as the first launch market. Cohere will supply its North agent platform, models, and search and retrieval capabilities; PwC will handle use-case selection, risk and regulatory assessment, workflow redesign, integrations, and scaled implementation. Clients can choose private cloud, on-premises, or air-gapped deployments. The commercial test remains ahead: the partners plan to expand successful use cases and evaluate productivity, cost, service, and decision-making outcomes, but have disclosed no results yet.
The companies announced the alliance on October 5, 2026, beginning with enterprise deployments in Canada.
Stated application areas include enterprise search, knowledge access, deep research, content generation, decision support, and task automation.
PwC will help determine where human judgment remains essential and establish performance, accountability, and controls for redesigned workflows.
Cohere and PwC announced a global enterprise AI alliance on October 5, 2026, with an initial launch in Canada. The partners aim to move organizations beyond isolated pilots by combining Cohere’s AI technology with PwC’s implementation and risk expertise, while helping clients retain control over sensitive data and the infrastructure running their systems.
PwC frames the partnership around an adoption problem: organizations have experimented with AI but struggle to turn separate projects into wider use with measurable results. Its release identifies several hurdles, including choosing valuable tasks, connecting AI to trusted business data, and navigating privacy, cybersecurity and regulatory requirements.
Domenic Marino, PwC Canada’s senior partner and CEO, said realizing AI’s potential requires more than technology. He presented the combination of implementation, industry and risk expertise as particularly relevant to regulated sectors, where trust and control over data are central concerns.
Cohere will provide North, its platform for AI agents, alongside enterprise models and search and information-retrieval capabilities. PwC’s responsibilities span selecting use cases, assessing risks and regulatory requirements, redesigning workflows, connecting technology to enterprise systems, and supporting implementation at scale.
The planned integration is meant to ground AI outputs in each client’s own knowledge and business context. The companies identify immediate applications across enterprise search, knowledge access, deep research, content generation, decision support and task automation—not just a single departmental tool.
The alliance includes three deployment options: private cloud, on-premises systems and isolated, air-gapped environments. The companies say these choices will give organizations greater control over sensitive data and infrastructure while expanding AI use across business workflows.
Annie Veillet, PwC Canada’s technology leader, said trust requires AI to operate securely inside an organization’s technology and data environment. She called for clear governance, explainability and human oversight from the outset, pairing deployment flexibility with implementation and risk management rather than treating hosting choices as the whole solution.
The partners’ stated next step is to scale successful use cases over time and measure results against business priorities. PwC names productivity, costs, service delivery and decision-making as those measures. That makes measurable business value an objective of the alliance, not a result already demonstrated by its announcement.
Loading discussion...
Join the conversation
Explain where outside expertise helps—and where employees should have the final say.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.