Coveo — Enterprise AI-Relevance Platform for Intelligent Search

AI-powered enterprise search platform provider offering unified content indexing, intelligent search, and generative experiences across commerce, service, websites, and workplaces.

Overview

Founded in 2005 as a spin-off of Copernic Technologies by Laurent Simoneau, Richard Tessier, and Marc Sanfaçon, Coveo develops SaaS-based enterprise intelligent search solutions from Montreal and Quebec City. The publicly-traded company (TSX/NYSE: CVO) posted its highest new business bookings in company history in Q3 FY2026 (quarter ended December 31, 2025), with 13% year-over-year SaaS Subscription Revenue growth to $36.6M and generative AI solutions accounting for more than 25% of new bookings.

That generative AI traction reflects a deliberate repositioning. By February 2026, Coveo had launched a Hosted MCP Server connecting its AI-Relevance Platform directly to ChatGPT Enterprise and Anthropic's Claude. This embeds its retrieval layer into third-party agentic stacks rather than competing as a standalone search destination. The pricing decision reinforces the strategy: no separate MCP fee; queries fold into existing consumption-based licensing. Ten customers were already using the hosted MCP server in production at launch.

In March 2026, Coveo extended its Canadian public-sector push further by announcing a sovereign AI partnership with Bell Canada, integrating the AI-Relevance Platform into Bell AI Fabric and targeting federal and provincial governments alongside regulated industries. Bell's Ateko division handles solution design, integration, and governance advisory. The partnership builds on the non-binding Memorandum of Understanding Coveo signed with the Government of Canada in December 2025, which carries execution risk as a non-binding instrument with no disclosed milestones or timelines.

Enterprise validation has followed. SAP's deployment across 11.2 million documents from 47 sources achieved 80% self-service success rates and €186 million in cost avoidance across 300 million cloud subscribers. Manulife's implementation in Global Wealth & Asset Management achieved 99.6% accuracy and doubled digital engagement. Executive Chairman Louis Têtu was appointed to Canada's AI Strategy Task Force in October 2025.

The Q3 FY2026 net loss of $7.2M (versus net income of $4.0M in Q3 FY2025) is largely an FX story: a $6.5M foreign exchange gain in Q3 FY2025 became a $1.2M loss in Q3 FY2026, accounting for $7.7M of the $11.2M swing. Adjusted EBITDA moved only $0.8M in the wrong direction. The company guides to breakeven EBITDA and positive operating cash flows for full fiscal year FY2026, though $14.3M in share buybacks year-to-date, executed while running near breakeven, is the more substantive financial question. No independent benchmark data from Gartner, Forrester, or IDC appeared in this period's sources to corroborate Coveo's competitive positioning claims against alternatives such as Glean, Elastic, or Microsoft Copilot connectors.

How Coveo Processes Documents

Coveo's AI-Relevance Platform consolidates content from 55+ sources, including Salesforce, ServiceNow, SAP, AWS, and Shopify, into a single AI-enriched index without requiring data migration. The platform combines proprietary relevance models with external large language models, allowing organizations to swap or layer LLMs without rebuilding the retrieval layer.

Machine learning runs through Coveo ML, which includes Intelligent Term Detection (ITD) and automatic relevance tuning that adjusts rankings based on behavioral signals without manual rule-writing. For generative answering, the platform's Relevance Generative Answering grounds responses in permission-aware content, meaning users only receive answers drawn from documents they are authorized to access. Document-level access controls enforce this at query time rather than at index time, a distinction that matters for regulated industries where row-level security is a compliance requirement.

The ML infrastructure runs on Netflix's Metaflow platform for model orchestration. The platform maintained 99.9%+ uptime including zero downtime during AWS outages, operating as a multi-tenant cloud service.

The February 2026 Hosted MCP Server extends this architecture outward: AI agents built on ChatGPT Enterprise or Claude can now query Coveo's unified index directly via the Model Context Protocol, without custom connectors. Coveo also introduced a RAG-as-a-Service layer for AWS agentic AI services, including Amazon Bedrock AgentCore, Amazon Bedrock Agents, and Amazon Q. This grounds those agents in organizational knowledge through the same platform. A Coveo app for ChatGPT allows business users to query enterprise data in natural language, listed in ChatGPT Enterprise's official Apps and Connectors directory.

The Bell Canada sovereign AI partnership adds a data-residency layer on top of this architecture. By embedding the platform within Bell AI Fabric, Canadian government and regulated-industry customers get data localization and governance built into the deployment model rather than configured after the fact. As Laurent Simoneau, CEO of Coveo, stated at the March 2026 announcement: "By embedding the Coveo AI-Relevance Platform within Bell AI Fabric, we are helping governments and enterprises to deploy advanced AI search and generative capabilities at scale with confidence to drive transformative outcomes while maintaining high standards of security, transparency and sovereignty."

Developers access the platform through a no-code builder or full API and SDK access. Builders and end-users interact with the same underlying index, keeping relevance tuning consistent across surfaces.

Use Cases

Enterprise Customer Support

SAP's deployment across 11.2 million documents from 47 sources is the clearest public benchmark of Coveo's scale. The implementation achieved €186 million in total cost avoidance across SAP's 300 million cloud subscribers, with €8 million in annual cost avoidance in SAP Concur alone. Relevance Generative Answering reduced SAP Concur case submissions by 30%, and the overall deployment reached an 80% self-service success rate. Coveo received the SAP Global Customer Experience Award for Partner Excellence 2025 recognizing this collaboration. New enterprise support customers in Q3 FY2026 include Boston Scientific, Tyler Technologies, and Roper St. Francis Healthcare, all in regulated, document-heavy verticals where permission-aware retrieval matters more than raw LLM capability.

E-Commerce Search and Recommendations

Commerce bookings represented nearly half of all new business in Q3 FY2026, a record quarter for the segment driven in part by the SAP partnership opening mid-market and international accounts. New Commerce customers include Insight Enterprises, Total Tools, La Vie en Rose, Metcash, Hugendubel Digital GmbH & Co., Tarkett S.A., and Shorr Packaging Corp, spanning distribution, retail, and manufacturing. Coveo's research found 70% of consumers accept generative AI shopping guidance, though only 3% begin a shopping journey with an AI assistant, suggesting the commerce opportunity lies in mid-funnel assistance rather than top-of-funnel discovery. The concentration risk is real: if SAP partnership momentum slows, commerce bookings could decelerate materially.

Financial Services Knowledge Management

Manulife's deployment in Global Wealth & Asset Management achieved 99.6% accuracy and doubled digital engagement, integrated with a generative AI-powered Wealth Advisor Assistant. The implementation won Forrester's 2025 Enterprise Architecture Award for North America. Deloitte Digital and Wind River Systems joined as generative AI customers in Q3 FY2026, extending Coveo's presence in professional services and industrial software. In Q2 FY2026, Nvidia, Intel, General Electric, and ABB Ltd. expanded existing generative AI usage on the platform.

Government and Sovereign AI

The Bell Canada partnership announced March 10, 2026 is Coveo's most concrete public-sector move to date. By integrating the AI-Relevance Platform into Bell AI Fabric, the two companies are targeting Canadian federal and provincial governments and regulated industries where data localization is a procurement requirement, not a preference. Bell's Ateko division handles integration and governance advisory, reducing implementation friction for large government deployments. John Watson, Group President at Bell, described the rationale at the announcement: "Canada has a responsibility and opportunity to lead in sovereign AI. Through our strategic partnership with Coveo, we are strengthening a made-in-Canada AI ecosystem that combines our secure national infrastructure and Bell AI Fabric with Coveo's world-class, internationally recognized AI-Relevance platform." The December 2025 MOU with the Government of Canada remains non-binding, with no implementation milestones or timelines disclosed publicly.

Technical Specifications

Q3 FY2026 Financial Snapshot (quarter ended December 31, 2025; unaudited IFRS):

Source: Coveo Q3 FY2026 earnings release.

Q4 FY2026 and Full-Year Guidance:

Note: Qubit Platform revenue fell 30% YoY to $0.8M in Q3 FY2026 and reaches zero in Q4, completing a deliberate wind-down. Core Platform growth of 15% excludes this drag. Full deprecation in Q4 FY2026 will provide cleaner visibility into core platform momentum going forward.

Resources

• Website
• AI Search Platform
• Technology Overview
• Q3 FY2026 Earnings Release: January 29, 2026
• Q2 FY2026 Earnings Release: November 2025
• Hosted MCP Server Announcement: February 10, 2026
• Bell Canada Sovereign AI Partnership: March 10, 2026
• SAP Cost Avoidance Case Study: Forbes
• Manulife / Forrester Enterprise Architecture Award
• Canada AI Strategy Task Force Appointment

Company Information

Headquarters: Montreal, Quebec, Canada

Founded: 2005

Founders: Laurent Simoneau (Co-Founder and CEO), Richard Tessier, Marc Sanfaçon

Executive Chairman: Louis Têtu

Chief Marketing Officer: Pranshu Tewari (appointed November 10, 2025; previously CMO at Mendix, acquired by Siemens)

Status: Publicly traded (TSX: CVO / NYSE: CVO)