Lenovo takes customer satisfaction to new heights with e-services enhanced by generative AI

Solving customer issues quickly, accurately and effectively on a global scale with new GenAI tools powered by GPT-4

Lenovo’s AI solution helps human experts improve the speed, accuracy and quality of interactions with customers on global support forums. The technology also enhances two existing customer-facing chatbots, enabling them to offer more human-like interactions and help with a wider range of support issues.

Challenge: Providing exemplary customer service

As customers, if we encounter an issue with our laptop, smartphone or tablet device, we naturally expect the vendor to help us resolve our problem quickly. Any company that fails to provide fast, friendly and effective support risks losing our loyalty as well as potential reputational damage, especially if we feel sufficiently frustrated to complain on social media.

We take e-services and customer support very seriously at Lenovo and Motorola, and we’re proud of the work we have done to build Lenovo Forum. This is a space where customers can post about the technical challenges they face and get rapid feedback from Lenovo experts. With two million users, 53,000 daily visits and hundreds of new topics posted regularly, responding promptly to users in the nine languages supported on the forum is labor-and cost-intensive—and we’re not alone in facing this challenge.

brand image - Colleagues on a production line discussing project.

Pioneering a new approach

To take customer support to the next level, Lenovo and Motorola turned to recent advances in generative AI (GenAI) for inspiration.

“GenAI is incredibly powerful,” says Rogerio Fragale, Global Customer Satisfaction E-services Director at Motorola. “We recognized that it could help service agents understand, translate and generate responses to customer queries faster and more accurately. Moreover, it could also give our existing intent-driven, customer-facing chatbots the ability to manage an even broader range of queries.”

Because of its sheer versatility, GenAI also presents fresh challenges. GenAI foundational models ingest massive volumes of data from the open internet, which means that their responses can be unpredictable and, on occasion, include inappropriate or even offensive material. Preventing this risk requires robust controls, careful development and rigorous testing.


2 million+ Lenovo Forum users

9 languages used to communicate customer support

50% of young consumers feel comfortable using AI support1


Solution: Lenovo powers Lenovo

Taking a phased approach to GenAI development, Lenovo built its new Lenovo Forum assistant using the GPT-4 Large Language Model (LLM) from OpenAI and Microsoft Azure cloud services. To ensure that the assistant could provide relevant, accurate and detailed answers to customer queries, the Lenovo team trained it using existing forum posts and information from Lenovo’s internal knowledge center.

Today, the assistant provides human service agents with automatic translations of forum posts written in up to nine languages, the ability to extract, translate and summarize information from images, as well as the capability to generate forum post responses with the click of a button. After checking autogenerated suggestions, service agents post responses to the forum and the AI assistant automatically translates them back into the customer’s preferred language.

SSG AI Pillars graphic: Security, People, Technology, Processes

Four Pillars of AI Readiness

Lenovo made sure that every stage of development aligned with the company’s four pillars of AI-preparedness: security, people, technology, and processes.

Security: Lenovo ensures that any sensitive customer data—whether it was shared via forum posts, images or as part of a chatbot conversation—is protected with the highest levels of security, and in compliance with privacy regulations.

People: Service agents make sure that GenAI forum responses are accurate, understandable and effective before they are posted online or to internal systems.

Technology: To mitigate the risk of inappropriate responses, Lenovo trained the assistant using only vetted Lenovo data, making out-of-scope responses practically impossible.

Process: To ensure safety, security and service quality, Lenovo put the GenAI assistant through a rigorous set of pre-production tests and constantly monitors the live solution for bugs or errors.

brand image - A diverse group of young professionals working on a project in a conference room.

Making chatbots even more capable

The second phase of Lenovo’s e-services transformation journey focused on enhancing the company’s existing customer-facing chatbots: Moli (for Motorola customers) and Lena (for Lenovo users).

Lena and Moli are already helping to provide fast responses for common, easy-to-fix customer issues. Our aim is that GenAI will be able help Lena and Moli to offer a more human-like interaction by reducing the frequency of intent validation or ‘did-you-mean-this?’ requests. GenAI will also make it possible for the chatbots to understand and formulate meaningful responses when customers ask them multiple questions simultaneously. Ultimately, this will make it possible for Lena and Moli to help with a wider range of support issues and reduce the workloads of our human support teams.

Result: Boosting the productivity and efficiency of e-services

Even though Lenovo’s GenAI forum assistant has been live for less than a year, the e-services team is already seeing significant benefits. So far, the forum assistant can automatically generate responses to customer queries with an 82% accuracy rate—a figure that will keep increasing as human service agents train the assistant by verifying its autogenerated responses.

Peter Buco, EMEA eService Delivery Lead / Services Operations Sr Professional at Lenovo, says: “Within a year, we expect that we will be able to manage many more customer interactions via the Lenovo Forum, delivering 50% greater efficiency. For service agents, this means they will spend more time adding value to customer relationships and our customers will enjoy faster, more responsive forum support around-the-clock.”

He adds: “So far, we have compiled and published 25,000 answers using our AI solution. We’re seeing great scores around resolution rates and customer satisfaction: a 30% improvement in our top-three-box score and a 17% boost to overall satisfaction.”


82% accuracy rate achieved for AI-generated forum responses

50% efficiency improvement anticipated in one year

30% improvement in top-three-box satisfaction scores since deploying GenAI for the Lenovo Forum

>80% average top-three-box score for GenAI forum responses


Bringing new benefits to Lenovo customers

Lenovo is already exploring how it can transform its innovations in GenAI into new service offerings for business customers. When the enhanced versions of Lena and Moli go live, we will demonstrate that chatbots powered by GenAI are suitable and highly effective for enterprise use cases. In the years ahead, Lenovo sees that GenAI will become a vital tool for delivering e-services as well as pre- and post-sales support on a global scale.

Fragale adds: “Eary adoption of GenAI can give companies a competitive advantage. By developing the Lenovo Forum assistant, we have already seen dramatic cost-reductions and service improvements. We are excited to share this potential with our business customers and help them take the next big step in their GenAI journey.”

To find out how you could benefit from Lenovo’s experience and AI-powered solutions to improve your customer communities and chatbots, visit Lenovo.com or speak to your Lenovo contact.

 

1 Forrester, “Predictions 2024: Exploration Generates Progress”; Gartner, “Market Guide for Digital Customer Service and Support”

[ssba]

Lenovo powers Lenovo

As a global technology leader, Lenovo is pioneering the use of groundbreaking AI to enhance our own operations, so we can help enterprises transform and embrace smarter AI for all.

Learn more about our AI-powered transformation.
Don't Miss StoryHub Updates: