How does Lenovo's Knowledge Super Agent cut search time by 30%?

Lenovo built an internal AI agent that answers work questions using the company's own documents. It now serves 3,000 employees and returns an estimated 120 hours per person each year.

Knowledge was scattered across thousands of locations

Critical information lived in shared drives and intranet sites. It also sat in process portals. Business hubs and databases held more. Employees searched across multiple systems. They reviewed documents or asked colleagues and subject-matter experts for help. New employees decoded acronyms and policies. They also learned processes. Sellers hunted for validated information and positioning for customer work. Operations staff retrieved SOPs and troubleshooting guidance. HR and compliance teams accessed approved policies and regulatory guidance.

Sarah Lundgren, Director of Technical Enablement for Hybrid Cloud and AI at Lenovo, said employees were effectively spending a full working day each week tracking down information. Searching across systems. Asking colleagues. Waiting for answers.

From AIESA to Knowledge Super Agent

Lenovo initially created the AI Enterprise Solution Assistant, or AIESA, as an AI agent grounded in its own enterprise knowledge. That internal forerunner became the foundation for what evolved into Lenovo Knowledge Super Agent. The company studied where knowledge gaps created the most friction. Teams mapped employees' day-to-day needs. They prioritized high-value questions and workflows. They also tested prototypes with real users.

Michaela Hric, Director of WW SSG Services Transformation Office at Lenovo, said the experience had to feel effortless for employees even though behind the scenes they were solving for trusted data and permissions. Governance and continuous improvement were also part of the work. The solution had to retrieve the right information from multiple sources. It had to ground its answers in trusted content. It had to respect each user's permissions. It had to remain reliable as knowledge changed.

Permissions and citations keep answers grounded

Role-based permissions protect sensitive information. Source-backed answers help employees understand where responses came from. More than 85% of answers now include supporting citations. Users can flag weak responses and route issues to the appropriate content owners. They can also reach product or technical owners. The underlying AI foundation was designed to be flexible so teams could add knowledge sources and capabilities without rebuilding the solution from scratch. Training and change-management programs helped people make the agent part of their everyday work. Feedback loops let users flag weak responses. Issues route to the appropriate content, product or technical owners.

Adoption grew 60-fold

Deployment expanded to approximately 3,000 Lenovo employees. Across that deployment Knowledge Super Agent reduced time spent on knowledge-retrieval tasks by 30%. Escalations to subject-matter experts fell by 35%. Sellers prepared customer proposals twice as fast. The impact extended beyond search. Specialists gained more time for complex work. New employees found policies faster. Operations staff retrieved SOPs without delay.

The product now sells to other companies

By building and deploying the solution inside its own business, Lenovo gained practical experience. Refining it over time added more lessons. It learned to connect enterprise knowledge and govern access. It improved response quality and drove adoption at scale. Those lessons now inform the customer-ready Lenovo Knowledge Super Agent. Organizations can tailor it to their own information and workflows. Governance requirements can also be customized. It is available through the Lenovo AI Library and supported by Lenovo Hybrid AI Advantage. The announcement does not say how many organizations have purchased it.