Technology company HP is expanding its strategic partnership with OpenAI following a series of internal pilot projects, and plans to deploy AI across customer experience, partner services, software engineering, cybersecurity and employee productivity.
The initiative centers on OpenAI’s Frontier platform, which provides a framework for enterprises to deploy, govern and manage AI agents. HP will use the platform to manage AI workflows across its global operations, with controls over access, permissions and evaluation as it rolls out agents across the business.
Pilot Programs Inform Broader Rollout
HP began testing OpenAI Frontier in February 2026 across several business units to evaluate how AI could improve software development and operational efficiency.
OpenAI cited security as an example of effective automation in the pilot:
“One engineer used OpenAI models to move through 122 pull requests across 43 projects in a matter of weeks. A security team used these models to remediate several software bugs in a day, work they estimated could otherwise have taken up to a month.”
HP also said a security team used the technology to remediate several software vulnerabilities in a single day, compared with an estimated timeline of up to one month using previous processes.
The pilots also involved the use of ChatGPT and Codex to support research, software development and workflow automation. OpenAI said the results indicated opportunities to reduce manual effort and shorten development cycles across multiple teams.
Customer and Partner Services Among Initial Priorities
As it expands the program, HP plans to apply AI across customer and partner-facing services.
The company said it intends to introduce AI-powered capabilities across its online store, customer support channels and partner portal. More than 80% of HP’s business is conducted through channel partners, with over 100,000 partners using its Partner Portal globally.
“Frontier will help HP create a more consistent self-service layer across store, partner, chat, and voice experiences, giving customers and partners faster ways to get answers, complete routine workflows, and move toward resolution or conversion,” OpenAI stated.
For partners, planned use cases include AI agents that can provide always-on guidance across program navigation, business information and various aspects of partner operations management. The aim is to improve response times while reducing the manual workload for support teams.
HP is also exploring the use of AI within its Workforce Experience Platform (WXP), where agents could analyze device telemetry, support documentation and operational data to help identify issues such as application crashes, Wi-Fi connectivity problems and device performance.
The expanded partnership also signals HP’s intention to make greater use of frontier AI models as part of its enterprise AI strategy.
Unlike earlier generations of GenAI, frontier models are designed to perform more advanced reasoning, software engineering, planning and multi-step task execution. These capabilities make them suitable for AI agents that can interact with multiple enterprise systems, use organizational context and complete increasingly sophisticated workflows.
For customer experience teams, this could enable AI systems that handle more complex service requests, retrieve information from multiple business applications and provide more contextual responses across customer and partner interactions.
HP said Frontier will act as the orchestration layer for these deployments by connecting AI models with enterprise data, permissions and evaluation processes.
Governance And Security Remain Central to Enterprise Adoption
As enterprises expand the use of AI agents and frontier models become capable of autonomously finding and exploiting security vulnerabilities, security, governance and regulatory compliance are becoming primary concerns.
Enterprise AI systems often require access to customer information, internal documentation and business applications that they risk exposing without appropriate controls.
The need for governance extends beyond data access. Recent research from AI research non-profit Aithos found that frontier AI models can still produce unexpected results under testing, violating key provisions under legislation such as General Data Protection Regulation (GDPR) or the EU AI Act. As Nadia Kadhim, Executive Director at Aithos, told CX Today in a recent interview:
“I may have expected a couple of the weaker models or the less used models to perform worse than the… frontier models, as we call them, and for the frontier models to mostly hold the line. But that’s not what happened.”
The findings indicate that organizations adopting frontier AI should not assume that more capable models inherently reduce compliance or security risks. Instead, enterprises need governance frameworks that include access controls, output evaluation, audit trails and human oversight to meet regulatory and internal compliance requirements.
Frontier AI also makes it “easier, faster and cheaper for attackers to discover and exploit weaknesses,” warns the U.K.’s National Cyber Security Centre. “This means that tasks which once required specialist skills—such as writing exploit code, understanding system architecture or using attack tools—can increasingly be automated using AI.”
However, the center acknowledges that “AI is not only benefiting attackers, but can also significantly improve and accelerate cyber defense,” adding, “Organizations that adopt AI securely, while maintaining strong cyber security fundamentals, will be better placed to take advantage of these defensive capabilities.”
OpenAI said Frontier is designed to address these challenges through permission management, deployment controls and evaluation frameworks that govern how AI systems access information and interact with enterprise systems. The platform is also intended to provide visibility into AI usage across the organization, allowing teams to monitor performance, review outputs and support compliance requirements.
HP’s cybersecurity teams have used ChatGPT to accelerate vulnerability analysis and remediation, estimating that these deployments have freed approximately 82 hours of security team capacity each week, although broader adoption will depend on maintaining appropriate governance as AI systems become more deeply integrated into business operations.
The partnership indicates how enterprise AI strategies are evolving beyond standalone productivity tools towards integrated platforms that combine AI models with governance, security and operational oversight.
Enterprises are increasingly seeking platforms that can manage AI deployments across multiple business functions while maintaining consistent controls over data access and permissions, establishing frameworks that support customer service, partner engagement and operational efficiency.
