Lyra

Customer Success on autopilot

Comment

Inspiration

At my previous company, client onboarding was a constant struggle. Our growth team was buried in admin busywork instead of focusing on strategic initiatives. During that time I met Stiven, who faced the same onboarding headache at his company. We realized this was a widespread issue affecting multiple SaaS businesses, and teamed to solve it with AI.

We’re set to disrupt the $6 billion onboarding software market—growing at 20% CAGR—by revolutionizing customer success through AI. We’re starting with onboarding, helping teams overcome the very challenges we face.

What it does

Lyra is an AI-powered assistant that automates and personalizes the customer onboarding process for B2B SaaS companies. It replaces rigid project management workflows by autonomously planning and executing tasks for Customer Success teams—all within familiar tools like Slack and email. Lyra handles routine administrative work, such as:

  • Scheduling meetings and organizing calendar events
  • Communicating with clients through personalized emails and messages
  • Updating internal systems and trackers in real-time
  • Providing 24/7 AI-driven co-browsing assistance to clients

By offloading tedious tasks to Lyra, CSMs can focus on strategic initiatives, reducing time to value and decreasing client churn.

How we built it

  • AI-Powered Core: Developed using TypeScript and Node.js, integrating OpenAI's GPT-4 for advanced natural language processing to identify clients, tasks, and automate decision-making.
  • Graph Database & Vector Search: Utilized Neo4j for complex relationship management and implemented semantic search using OpenAI's embedding model for accurate task matching.
  • Slack Integration: Built a robust Slack integration using the Bolt framework to handle real-time message events and user interactions, serving as the primary interface for CSMs.
  • Adaptive Onboarding Playbook: Created an AI-driven system to generate and manage customized onboarding playbooks, automatically adjusting tasks based on client interactions and information.
  • Google Sheets Integration: Implemented automatic creation and updating of client-specific sheets, providing easy task tracking and visibility for both CSMs and clients.

Challenges we ran into

One of the main challenges was optimizing the language models to handle diverse and nuanced onboarding situations. We conducted extensive evaluations to improve the AI’s performance, which involved significant trial and error. Ensuring that Lyra could personalize interactions for different clients required fine-tuning the models and implementing feedback mechanisms. Integrating advanced AI into existing workflows without disrupting them also posed technical hurdles that we overcame.

Accomplishments that we're proud of

  • Functional Prototype: Successfully built a working prototype of Lyra within a limited timeframe.
  • AI Integration: Effectively integrated advanced AI technologies beyond LLMs like GraphRAG to enhance personalization and adaptability.
  • Seamless User Experience: Created an intuitive user interface within Slack and email, lowering the learning curve for users.
  • Positive Feedback: Received positive responses after initial demonstration to CSMs & prospective user, who expressed desire to adopt the product, demonstrations to CSMs and potential users, validating our solution.

What we learned

  • User-Centric Design is Crucial: The importance of building Lyra to align with current needs and workflows of CSMs.
  • Continuous Improvement: The necessity of iterative testing and refinement in developing effective AI models.
  • Integration Challenges: Gained insights into the complexities of integrating AI tools into existing systems and the importance of seamless integration.
  • Market Potential: Deepened our understanding of the significant demand for automation in customer success processes.

What's next for Lyra

  • Expanded Capabilities: Enhancing Lyra to cover the entire customer success journey, including activation, proactive churn prevention, and renewals.
  • Pilot Programs: Launching pilot programs with interested SaaS companies to gather real-world feedback and iterate on the product.
  • Refinement of AI Models: Continuously improving the AI for better accuracy and personalization.
  • Scaling and Partnerships: Exploring partnerships and scaling opportunities to bring Lyra to a broader market.
  • User Feedback Integration: Implementing features based on user suggestions to make Lyra even more valuable to CSMs.
Share this project:

Updates