AI coaching to drive organizational performance

Introduction

Welcome and Introduction

Hello there and welcome! I'm Shane Mallister, and this is another episode of MongoDB TV Cloud Connect. Cloud Connect streams most Thursdays and dives into what's good and exciting in the world of all things developer-related. Today's episode is no different. In this episode, we have not one but two guests, Thiago and Mia from Patron.ai, and we'll be delving into the world of AI coaching to drive organizational performance.

Housekeeping

Before we get started, here's some housekeeping:

• Drop a shout out in the chat either on LinkedIn or YouTube telling us who you are and where you're joining from.
• Use the chat in YouTube and LinkedIn to ask questions.
• If you're a seasoned viewer, welcome back! If you're new, a warm welcome to you.
• Past shows are available on LinkedIn and YouTube.
• The live stream is being recorded, and will be available on MongoDB's YouTube channel later.
• Please subscribe to the MongoDB channel on YouTube and follow us on LinkedIn for the latest news.

Last week in New York, MongoDB launched its MongoDB AI applications program and showcased features in MongoDB 8.0 during our largest local event on the calendar. If you're interested in learning more, visit mongodb.com to see what's new.

AI Coaching and Patron.ai

Today, we dive into a fascinating topic that's transforming businesses globally: AI-enabled coaching. We’re joined by Thiago and Mia from Patron.ai.

Thiago's Journey

Thiago's career path has been diverse, starting early in coding, leading to a background in IT management and computer science. His initial work involved research in NLP (Natural Language Processing) for genetic sequence reading. This early interest in machine learning led him to Finland in 2014, where he worked with startups focusing on NLP and chatbots. In 2019, he met Patron.ai's founders at a hackathon, joining them to use AI in transforming corporate training into scalable coaching.

Mia's Journey

Mia moved from Vietnam to Finland for her studies. She grew interested in NLP during an internship in Sweden. By 2019, she was deeply involved in consulting work related to chatbots, leading her to join Patron.ai as an intern. Her role has evolved into a technical product manager, involved in both NLP work and backend engineering.

Overview of Patron.ai

Founded initially as Panda Training, Patron.ai first evaluated corporate training efficiency. They discovered that short-term retention was significant, but lasting behavioral change was challenging. The pivot to AI coaching aimed to solve this, leveraging the scalable nature of AI coupled with the chatbot’s autonomy, essentially creating a coach always available for employees. The proprietary AI could tailor coaching programs based on each organization’s unique vision and goals.

Practical Usage of Patron.ai

Patron.ai provides interactive sessions via familiar tools like Microsoft Teams and Slack. These sessions are scheduled, often weekly, and each lasts around 30 minutes. The AI, named Panda, conducts these sessions, guiding employees through organizational and personal development questions.

Technology Stack

• Languages and Frameworks: Node.js, Flutter, and Python for side projects.
• Database: MongoDB Atlas due to its document model, ideal for managing conversational data.
• AI Models: Integration with OpenAI's models, benchmarking, and evaluation of others for optimal performance.

Coaching Sessions

Patients’ interactions are governed by prompt engineering, reducing hallucination or off-topic responses. Feedback mechanisms embedded at the end of sessions help improve the AI's effectiveness.

Adaptation during COVID-19

COVID-19 brought unprecedented changes to workplace dynamics, enhancing the need for adaptive AI coaches like Panda. Patron.ai’s AI coaching became crucial in supporting remote employees' work-life balance.

Future Directions

As AI continues to evolve, Patron.ai aims to integrate more sophisticated models, enhance response speeds, and support multilingual capabilities, particularly for less represented languages. The growing acceptance and expectation of AI-enabled tools drive the necessity for continued innovation and agility in service delivery.

Conclusion and Contact Information

Patron.ai represents a significant step forward in corporate coaching, offering a scalable, interactive AI system adaptable to individual and organizational needs. To learn more, visit patron.ai and follow them on LinkedIn.

Introduction

• AI coaching
• NLP
• Corporate training
• Behavioral change
• Patron.ai
• Chatbots
• MongoDB
• OpenAI
• Microsoft Teams
• Slack

Introduction

Q: How did Patron.ai originate? A: Patron.ai began as Panda Training, focusing on the effectiveness of corporate training. It pivoted to AI coaching in 2019 to address the need for scalable behavioral change in organizations.

Q: What technologies underpin Patron.ai? A: Patron.ai uses Node.js, Flutter, Python, and MongoDB Atlas. Their AI integrations include OpenAI models tailored for NLP and conversational interfaces.

Q: How does Patron.ai ensure the suitability of its AI models? A: Patron.ai regularly benchmarks and evaluates different AI models, including back-up solutions. They prioritize robustness and cost-effectiveness in their integrations.

Q: What benefits does AI coaching offer over traditional human coaching? A: AI coaching, particularly through Patron.ai, offers scalable, consistent, and always-available sessions. It provides personalized coaching without judgment and can gather valuable organizational insights through anonymous data.

Q: How does Patron.ai handle conversation data? A: Conversation data is anonymized and used for improving session quality and generating organizational insights. Patron.ai adheres to strict data security and compliance policies to ensure user privacy.

Q: Can Patron.ai be used in multiple languages? A: Currently, Patron.ai mainly supports English, but they are working on expanding support to more languages, including less common ones, to cater to a wider global audience.