Emma automates 88% of presales conversations with AI agents
By introducing Sprinklr’s multilingual AI agents, Emma streamlined presales inquiries, eased agent workload and delivered faster, more consistent customer guidance across 11 markets.
Challenge
Emma is a leading direct-to-consumer sleep brand offering mattresses, pillows and sleep accessories online across multiple global markets. Because customers often need guidance when choosing a mattress without seeing it in person, Emma places a strong emphasis on delivering clear, high-quality presales support that helps shoppers feel confident in their decisions.
The AI agent answers routine questions fast. Our customers feel the difference, and our agents can focus on the more complex work.”
As the company grew, Emma saw an opportunity to modernize its digital customer journey by improving how it handles repetitive presales inquiries across markets and languages. Emma's previous, rules-based chatbot did not allow for automation, which put ongoing pressure on agents to stay fully aligned with frequent product and promotional updates.
Emma wanted to:
- Streamline presales conversations
- Reduce operational burden on agents
- Provide more consistent and accurate answers
- Offer responsive, 24/7 support to customers across all markets
This created a clear opportunity to rethink how the company could use conversational AI to elevate the customer experience while driving stronger efficiency internally.
“We wanted to automate more customer inquiries while improving overall customer satisfaction,” explains Alexandra Pashkovska, Digital Transformation Manager at Emma.
Solution
AI‑powered presales guidance at scale
Emma implemented Sprinklr's scalable AI agent to provide fast, friendly and consistent presales guidance across 11 markets and multiple languages.
The AI agent is powered by Sprinklr’s Smart FAQ model, which uses retrieval- augmented generation to answer questions based on approved knowledge sources. Smart FAQ allows Emma to ground the bot’s responses in accurate, structured content, supporting consistent, trustworthy guidance across.
The AI agent now answers common questions and helps customers understand product differences. It delivers this support 24/7 — ensuring customers get reliable information at any time.
Visibility into customer journeys without compromising privacy
To better understand downstream customer behavior, Emma implemented a structured UTM tracking framework. Every product link shared by the virtual agent includes custom, market-specific UTM parameters mapped to Emma’s Google Analytics setup. This approach enables the team to observe actions such as product views and purchases, giving the team visibility into how customers move through their journeys without identifying individual customers.
Built‑in safeguards ensure trust and accuracy
The Emma team has also implemented clear governance controls, including rules that prevent the AI agent from providing medical advice. These safeguards — introduced after identifying occasional hallucinations early on — ensure customers receive accurate, trustworthy guidance across every market.
Together, these capabilities created a fully automated presales experience that is easier for Emma to maintain, lighter on agents and more responsive for customers.
Outcome
Emma surpassed its automation goals significantly. What began as a target of 30% automation grew to 88% automated presales conversations across 11 markets. This reflects major efficiency gains and reduced load on agents. And customers have responded positively, giving the virtual agent a CSAT score of 70 and a customer effort score of 74 — which are both higher than human-assisted interactions.
While sales outcomes were not the intention of the project, Emma noticed promising early indicators. Customers who engage with Sprinklr's AI agent are four times more likely to convert and to place higher value orders. This suggests that the AI agent may contribute to stronger buying intent in some markets.
Internally, Emma has significantly reduced the effort required to maintain accurate presales information, decreased dependency on agent training and improved support consistency. Externally, customers now enjoy immediate, reliable, always-on guidance — a meaningful evolution of the presales experience.
“The AI agent answers routine questions fast,” says Pashkovska. “Our customers feel the difference, and our agents can focus on the more complex work.”
