Reallysec Builds a Modern AI-Driven Search Platform for Cisco

Delivering a Modern AI-Driven Search Experience for Customers, Partners, and Employees

Cisco Systems, founded in 1984, is a global leader in networking technology, dedicated to providing critical technology infrastructure for the global digital economy. Today, over 87% of Fortune 500 companies rely on Cisco's networking hardware, software, communication equipment, and security solutions.

Within Cisco's global operations, content search capability has become a critical infrastructure component, serving core business functions including customer support, product portals, and technical documentation management. For example, in customer support scenarios, over 11,000 support engineers use the search platform daily to quickly retrieve solutions from millions of documents, handling more than 2 million service requests annually.

Meanwhile, the Cisco.com search system provides global visitors with the ability to find information across hundreds of thousands of web pages, data sheets, technical guides, and product resources, serving as a vital bridge connecting users with content.

Today, customers expect to find the information they need instantly. Accurate and efficient search is at the core of customer relationships and product awareness. — Sujith Joseph, Principal Enterprise Search and Cloud Architect, Cisco Systems.

Beyond accuracy, search response speed is equally critical. Even a 0.5-second delay can impact user click-through rates or technical support experience.

To address these challenges, Cisco launched an AI-driven enterprise search modernization initiative, leveraging cloud-native architecture and intelligent search engines to reshape its content access systems for both customers and employees. Prioritizing efficiency and controllability, the upgrade initially focused on customer support and Cisco.com search systems, with plans to gradually expand to internal Intranet and other scenarios.

Elastic at the Core of Cisco's New Search Platform

"Reallysec demonstrated to us that they could deliver a 100% cloud-native solution, including integration with Google Kubernetes Engine." — Sujith Joseph, Principal Enterprise Search and Cloud Architect, Cisco Systems

Elasticsearch, now deployed on Elastic Cloud within a Kubernetes environment, has become the core engine of Cisco's next-generation enterprise search architecture. Joseph specifically noted that the Re-imagined Topic Search tool for customer support scenarios achieved significant performance improvements, enabling support engineers to quickly retrieve documents highly relevant to service requests.

During support calls, engineers can search historical cases in real time based on error messages provided by customers. The system automatically matches and returns associated support records, product defects, knowledge base articles, and internal discussion content. With this intelligent search capability, Cisco saves support engineers over 5,000 hours per month, significantly improving support response efficiency and customer experience.

"The feedback from our engineers has been very positive. They now use topic search to resolve 90% of service requests. They can provide a better customer experience by easily finding the information they need and resolving issues faster than before." — Sujith Joseph, Principal Enterprise Search and Cloud Architect, Cisco Systems

Cisco.com Search Experience Upgrade: The AI-Driven Re-imagined Search Platform

The search functionality on Cisco's official website (Cisco.com) is powered by the newly built Re-imagined Search Platform, an AI-driven enterprise search solution built on Google Cloud services and Elasticsearch.

The platform uses Neural Question Generation technology to intelligently analyze Cisco.com page content, providing users with real-time auto-completion and question suggestions. Additionally, the system features hybrid semantic and textual search capabilities (Hybrid Semantic + Textual Search), enabling deep question-answering and precise positioning at the paragraph level within documents.

"We have the ability to pre-process search queries and direct search results to specific paragraphs within documents, not just the top of a web page. When users click on a search result, the system takes them directly to the most relevant content fragment — a significant user experience upgrade compared to the previous approach of only navigating to the full page." — Sujith Joseph

Joseph further noted that the combination of AI and Elasticsearch has greatly enriched the end-user search experience:

"We use deep learning models to semantically reconstruct and pre-process user input, then pass it to the Elastic engine for retrieval. This mechanism enables us to generate search result pages with higher relevance, allowing users to find the information they need faster and more accurately."

"When you combine Elastic's accuracy and speed with the power of Google Cloud, you can build a very stable and cost-effective search platform while providing users with a delightful experience." — Sujith Joseph

Boosting User Engagement and Significantly Optimizing Operational Costs

With the new-generation search platform, user click-through rates have increased substantially, and the search experience has become smoother and more natural. System response speed improved by up to 73%, significantly reducing the risk of users abandoning pages during loading.

By improving search efficiency and information matching accuracy, Cisco has not only enhanced user engagement but also significantly reduced support burden and resource consumption at the operational level, achieving dual value in experience improvement and cost optimization.

"Because Elastic integrates with the latest cloud technologies and platforms, we can innovate faster with Elastic. It has also improved our overall operational efficiency, enabling us to deliver highly accurate business information more quickly." — Sujith Joseph

Rapid Platform Expansion and Continuous AI Capability Iteration

Since the Re-imagined Search platform launched for customer support and Cisco.com, Cisco's search team has expanded the solution to over 50 internal and external applications, including critical systems such as the enterprise Intranet. The team continues to introduce more AI tools to enhance search intelligence and response efficiency across various scenarios.

Customer Support Scenarios

• Using AI models to automatically generate search queries from support tickets, sending queries in real time to Elastic to obtain highly relevant content recommendations and provide instant feedback to customers.
• By predicting the optimal resolution paths engineers have taken for similar tickets, helping support teams save diagnostic time and accelerate problem resolution.

Intranet Internal Search

• Embedding-based semantic search on employee knowledge base content, combined with AI models, to deliver precise Q&A and intelligent auto-completion experiences.
• Fully leveraging Elasticsearch's Dense Vector Search capabilities to improve the efficiency and accuracy of internal information retrieval.

"We have made significant progress in search. We use AI to make our Elastic-powered search platform better and relevant to a broader audience across the entire enterprise." — Sujith Joseph