The client is a major global beauty brand whose customer service agents respond to online ratings and reviews across several of its brands, working through the Bazaarvoice platform. Responding to reviews, whether positive or negative, is a key way the brand connects with customers, builds trust and loyalty and recommends alternative or complementary products, provided the responses are on-brand, helpful and authentic.

Whether online reviews are positive or negative, responding to them is a valuable way for a brand to connect with its customers. Responding to reviews and ratings builds trust and loyalty, but only if responses are on-brand, helpful and feel authentic. Effectively addressing the concerns raised in critical reviews, or recommending products to complement customer favorites, creates rapport and improves brand perception.

For this beauty brand, drafting those responses from scratch was time-consuming for agents. Brand voice was not always consistent across agent responses, and finding the right alternative or complementary products to recommend could be cumbersome. Agents also had to log in to Bazaarvoice separately for each brand and copy information between systems by hand.

When Laivly began working with the company, the baseline agent handle time per review case was 400 seconds, and agents were completing about nine reviews per hour. The brand wanted to create custom, personalized review responses faster and with more consistency, without losing the empathy and personal touch that make a response feel genuine.

Laivly deployed Sidd, its digital intelligence, to assist the brand's customer service agents by eliminating manual steps in their workflow so they could focus on personalization and empathy. Laivly Smart Response leverages generative AI together with proprietary language filters, and Sidd combines process automation, machine learning and generative AI to simplify the ratings and reviews workflow.

Sidd analyzes each review and then drafts a selection of on-brand responses, including alternative or complementary product recommendations when appropriate. Agents select the Sidd-generated response that reflects the right sentiment, modify or personalize it further to suit the situation, and reply to the customer. Sidd also brings the relevant information to the agent without copying and pasting, and logs into each brand's Bazaarvoice account on the agent's behalf.

The effect on throughput was immediate and grew with use. Agent handle time per review case fell from 400 seconds to 234 seconds. Productivity initially rose from 9 to 13 reviews per hour (RPH) and, with continued use, reached 15.4 RPH. After implementing Sidd, 3,422 cases were handled in 220 hours versus the 422 hours it previously took, which means that for every 500 hours staffed with agents supported by Sidd, an additional 4,001 cases can be resolved.

• 71% more reviews handled per hour, rising from 9 to 15.4 reviews per hour
• 40% reduction in handle time, from 400 seconds to 234 seconds per review case
• 48% fewer hours required for the same volume: 3,422 cases in 220 hours instead of 422
• 89% capacity increase for additional cases, or 4,001 extra cases per 500 staffed hours
• $48K net monthly savings
• Consistent, on-brand responses with product recommendations, without logging into Bazaarvoice for each brand