April 4th, 2024


Eric Ciarla imageEric Ciarla

Meet MongoDBs CoachGTM.ai

The Mendable bot saving MongoDBs customer success and sales teams millions per year

Meet MongoDBs CoachGTM.ai image

What is CoachGTM?

CoachGTM, a Mendable AI Slack bot powered by MongoDB Atlas Vector Search, equips MongoDB’s teams with the knowledge and expertise they need to engage with customers meaningfully, reducing the risk of churn and fostering lasting relationships.

CoachGTM is designed to utilize an extensive dataset, and provides support in a number of ways, including offering technical guidance, process assistance, and industry expertise. The ultimate aim of CoachGTM is to enhance customer success and improve sales problem-solving in an accessible way.

Business impact at MongoDB

CoachGTM answers thousands of questions per month and saves an estimated ~20 minutes per question asked. At MongoDB, this translates into an expected $2 million in savings annually by answering over 50,000 questions for ~1000 active users in customer success(CS) and sales. Indeed, CoachGTM has had an impact on a number of teams across MongoDB:

Customer success managers (CSMs):

CoachGTM provides CSMs with rapid access to accurate information, enabling quick responses to customer queries, especially when technical experts are unavailable. MongoDB CSMs appreciate the ability to provide immediate answers.

Sales representatives/account executives:

The platform helps identify customer issues quickly, enhancing deal closures. A sales rep noted CoachGTM’s role in speeding up problem resolution and value delivery during client meetings.

Sales engineers/solutions architects:

Routine inquiry handling time is reduced, as evidenced by a Solutions Architect who reported fewer interruptions since using CoachGTM, indicating a decrease in basic query volume.

Transforming sales

MongoDB’s partnership with Mendable and the implementation of CoachGTM is a testament to the transformative power of generative AI in enhancing business operations. As James Underhill, Director, Sales Operations and Strategy at MongoDB noted, “Mendable allows our sellers to get deeper into problem-solving faster, creating more value and trust in our customer interactions.”

The benefits of generative AI for sales and CS teams are vast, and getting in early means your team will have a huge leg up on the competition! Getting started is as simple as signing up at Mendable.ai and ingesting your data from one of the data sources we currently support such as Zendesk, Google Drive, and Salesforce. From here you can one click install a bot into your Slack or Teams workspace and start supercharging your teams.

In addition to answering questions, Mendable also provides a Tools & Actions feature which allows for tasks like updating CRMs, automatic prospect research, and AI-powered meeting prep all from the same interface. To get a walkthrough of the platform and all of its potential, book an AI consultation call with a Mendable expert!

An LLM stack powered by MongoDB Atlas

Mendable uses MongoDB Atlas Vector Search to power CoachGTM, which involves several key technical steps:

Indexing with MongoDB Atlas Vector Search:

CoachGTM leverages the power of MongoDB Atlas Vector Search to create highly efficient index vector representations of thousands of documents, including websites, pdfs, and docs. They are ingested to the index through the Mendable API every 24 hours so information is always up to date.

Scalability and performance optimization:

MongoDB Atlas Vector Search is designed to handle large-scale vector data with exceptional efficiency. Compared to other stacks that Mendable tried, Atlas Vector Search came out on top because it is able to seamlessly scale search infrastructure as data grows, ensuring optimal performance at all times.

Hybrid search implementation:

Mendable uses a hybrid search approach that combines keyword-based searches with vector similarity searches. In testing, hybrid search works better than semantic search alone, and boosted correct document matching by around 20%. The system performs both a keyword-based search and a vector similarity search simultaneously on MongoDB Atlas, integrating the results to generate a comprehensive list of documents.

Result fusion and ranking:

After the hybrid search, Mendable employs LLM-powered result fusion techniques to combine and rerank the search results from MongoDB Atlas that are most relevant to the query. By considering factors like relevance, freshness, and the end user’s profile, the system provides the most accurate and pertinent search results to users.

By harnessing the power of MongoDB Atlas Vector Search, CoachGTM delivers lightning-fast and highly relevant context to MongoDBs sellers, ushering in a new AI wave in revenue operations.