Tidio empowers e-commerce businesses to resolve up to 90% of customer inquiries automatically using their Claude-powered Lyro AI agent.
With Claude, Tidio:
- Achieved 71% automation of Tidioâs own customer support across live chat and email channels
- Grew Lyro adoption by 700% in just one year
- Resolved over 2 million customer conversations through AI automation
- Offers a risk-free guarantee: below 50% automation rate, customers pay nothing
How Tidio tackled repetitive customer support questions
Tidio, founded in 2015, started as a live chat solution but evolved into a comprehensive customer service platform focused on e-commerce. As they scaled, they observed support teams spent enormous time answering the same questions.
âLive chat is used when customers feel the issue is easyâlike login problems, system bugs, or discount code failures,â explained Marcin GwizdaÅa, CTO at Tidio. These inquiries required quick responses and consumed valuable agent time.
They faced different challenges for email support. GwizdaÅa noted, âIn e-commerce, customers mostly ask about issues requiring action. âI want to return, what should I do?ââ These interactions demanded deeper understanding and often needed to trigger specific business processes.
Why Tidio selected Claude to power their AI agent
When evaluating AI models, Tidio found Claude stood apart for several critical reasons:
âChatGPT wasnât a good solution for us. We needed something more sophisticated and developer-friendly that we could integrate into our system. Then we found Claude and fell in love,â said Tytus GoÅas, CEO of Tidio. âWe loved the quality, design, and attention to building a solution for SaaS creators.â
When selecting an AI partner, Tidioâs main concerns were safety and reliability. âClaude consistently follows instructions and adheres to prompts better than alternatives we tested,â explained GwizdaÅa. âIts built-in safety guardrails make it extremely difficult to jailbreak, and with our additional security layers, weâre confident our Lyro agent wonât generate harmful or inappropriate responses to customers.â
Claudeâs capabilities have helped Tidio stand out significantly in Poland, their local market. âWe did extensive testing of all solutions and really care about quality for our end users. We know they wonât be happy if our chatbot hallucinates or promises impossible outcomes,â added GwizdaÅa.
How Claude powers comprehensive customer service automation
In mid-2023, Tidio became the second company worldwide to release a Claude-powered conversational AI product. Their Lyro agent handles customer interactions across multiple channels:
For live chat, Lyro delivers responses within 5-20 seconds while accurately addressing common questionsâcritical for e-commerce engagement.
Claudeâs reasoning capabilities tackle complex issues requiring specific actions for email support. âFunction calling in Claude models helped understand constraints and make plans,â said GwizdaÅa.
Tidio has developed a system of agentsâdifferent Claude instances working together behind the scenes. âOne agent rates the conversation, another summarizes it, creating a network of specialized AI processes,â explained GwizdaÅa.
Tidio engineered a clever solution to address API capacity challenges. âWe implemented a dynamic routing system that automatically selects between native Anthropic API and Google Cloudâs Vertex AI based on performance metrics from the past 30 minutes, ensuring optimal response times.â
Transforming support operations and boosting revenue
Implementing Claude-powered automation has transformed both Tidioâs operations and their customersâ businesses.
Tidioâs support team achieved a 71% automation rate across all channels. âWe didnât reduce our workforce. Our support agents now have more ambitious tasks than answering for the 341st time why someone cannot log in,â noted GwizdaÅa.
Despite being built primarily for commerce businesses, Claudeâs flexibility allowed Tidio to achieve remarkable results on emailâa channel with more complex challenges than live chat. âEmails contain multi-paragraph explanations and require more nuanced understanding. The problems are more complicated and convoluted.â
Tidioâs Product Recommendation Engine uses Claude to analyze vague requests, search thousands of products, and make personalized recommendations that match subtle requirements for customers. One customer, Belasante, generated over $60,000 in additional revenue using this capability.
âPeople visit a website selling thousands of RC toys and ask âIâm looking for something for my eight-year-old niece.â With just these details, Claude can understand the request, search products, verify age-appropriateness, and recommend matches in six to nine seconds,â GwizdaÅa shared.
Building the future of customer service with Claude
Tidio aims for 80% automation rates and to expand into a comprehensive multi-channel help desk solution. âThe revolution happened. Now comes evolution,â said GwizdaÅa. âWe expect Claude models to improve, enabling more complex tasks.â
The company is developing self-learning loops where Claude analyzes past interactions, extracts valuable knowledge, and feeds it back into their knowledge base. âIf a question and answer look repeatable, we extract it as knowledge and put it into the agent to create a continuously improving system,â explained GwizdaÅa.
Beyond automation, Tidio sees AI as a tool for deeper customer understanding. âWe can help our customers understand their customers better, ensuring end users are happy with the services they choose.â As an early Claude adopter, Tidio remains committed to innovation in customer experience while maintaining their focus on quality, safety, and exceptional service.