Questflow redefines the future of work with AI with MongoDB Atlas

INTRODUCTION

Multi-agent AI Orchestration

The way people work is changing in a disruptive way. For hundreds of years, people have been at the center of every kind of office work, no matter how tedious. But in the future, office agents will collaborate with AI tools and assistants in many ways. We may even see “1+N” scenarios in which one person works with several AI agents. A significant amount of tedious work, like data entry, will be delegated to AI agents, while people will focus on high-value, essential tasks like coordination, approvals, and verification, according to Bob Xu, Chief Executive Officer of Questflow.

Questflow Labs, a startup supported by MiraclePlus, is a decentralized autonomous AI agent network that dispatches real-world incentives for autonomous AI agents. It allows users to orchestrate multiple AI agents to take action on autopilot, distribute real-world incentives to creators of AI agents, and collaborate with human agents. Compared with companies emphasizing traditional robotic process automation, Questflow provides a multi-agent AI orchestration framework that allows human intervention for task or transaction approval, making it a more efficient, next-generation tool for collaborative workflows.

Questflow currently supports mainly small and medium-sized startups with limited resources, enabling them to build their own AI agents and quickly automate their own workflows. These tasks could include collecting and analyzing business information, posting social content, replying to emails, and writing meeting minutes. Organizations and individuals can use MongoDB-powered Questflow tools to intelligently perform different tasks and services without developing their own automation systems.

THE CHALLENGE

Building personalized and dynamic AI agents

In the era of AI, having AI agents is now a trend. AI agents can perform repetitive, tedious, or data-intensive tasks such as data analysis, customer service, and document processing. And integrating AI agents into platforms like websites, mobile apps, or messaging apps allows them to provide real-time assistance and support.

Creating AI agents is one of Questflow’s flagship products. The company aims to make AI agents more agile, flexible, and efficient than humans.

To achieve these goals, it is crucial to design AI agents with diverse characteristics. For example, they should be able to constantly evolve and integrate new knowledge into their practical experience. The ability to personalize AI agents is also key, as each one should be equipped with unique memory capabilities, analytical skills, and coordination abilities.

To enhance the dynamism and personalization of AI agents, Questflow’s backend system handles a substantial amount of unstructured and AI data—including text, images, and audio—which is stored in vector databases for similarity search and data analysis. As AI agents become more intelligent, the demand for powerful database software to underpin them will only grow.

Questflow additionally specializes in leveraging natural language to automate tasks for AI agents. The execution of this process, however, involves multiple interconnected steps, including reasoning, understanding, and implementation. Previously, the coordination between these steps was inefficient due to limitations in AI models and data analysis capabilities. Therefore, Questflow required an effective solution to enhance the overall process, enabling AI agents to perform their tasks seamlessly.

To meet these higher demands, Questflow adopted MongoDB Atlas for its data management services.

THE SOLUTION

Enhancing customer experience with simplified data analysis

As a modern, multi-cloud database platform, MongoDB Atlas unifies operational, analytical, and generative AI data services to simplify the development of intelligent applications.

“To redefine the way people work in the future, we must first redefine our own work processes, which means enhancing the capabilities of our AI agents with exceptional data storage and processing capabilities,” said Carney Chu, Co-founder and Chief Technology Officer of Questflow. “MongoDB is highly adaptable, and MongoDB Atlas in particular is tailored for AI, which is a great fit for our business.”

Vector search, flexible scalability, and cloud deployment are among the most appreciated features of MongoDB Atlas for Questflow so far.

Questflow uses MongoDB Atlas Vector Search for all its vector data management services. Questflow’s products are designed in the form of conversational interfaces, helping users automatically resolve issues. This requires a large amount of vector data to be stored and processed in the backend. By storing vector data in MongoDB Atlas, customers are able to perform a vector search on the platform and deliver highly accurate generative AI content. MongoDB can support both traditional data and vector data of an embedded knowledge base. This allows customers to perform hybrid search locally without the need to develop software.

All AI companies face a common problem of increasing volume. MongoDB Atlas is able to automatically scale up both cluster size and storage size to accommodate Questflow’s expanding business needs after two years. MongoDB's flexible scaling capability is valuable for startups like Questflow, because it allows developers to focus on collecting and classifying business data without allocating time and effort for daily maintenance.

To enhance the cost control, convenience, and security management, Questflow chose to deploy MongoDB Atlas on Amazon Web Services (AWS), enabling a plug-and-play experience with just a few clicks and backend configuration.

THE RESULTS

Optimize innovation with next-generation AI technology

Questflow’s AI agents can enable new work modes with fewer resources. With MongoDB Atlas, Questflow has helped several startups to have more time and budgets for innovation. For example, iMindMap, an AI mind mapping company, automated its blog publishing service with Questflow. And OpenVC, an open source VC database company, automated its data retrieval, filtering, and updates through Questflow.

Xu hopes to go beyond the current model of chatting with AI. For example, if a user wants to create a poster, they can now chat with AI to create the poster, but modifications might need to be made on other platforms, which could lead to a disjointed user experience. Questflow’s next goal is to enable collaborative editing and optimization of the content generated by AI among multiple people.

Questflow believes that there will be increased collaboration between human agents and AI agents, opening the door to further innovation.

With this vision, Questflow is also providing automation services for large and medium-sized organizations, ranging from replying emails, operating social media accounts, and managing marketing activities. For example, a luxury brand company needs to rely on their staff to manually post pictures and articles of their new products on hundreds of social media accounts.. Questflow is looking at ways to provide a full suite of AI services to meet such demands, enabling a team of AI agents to collaboratively automate these tasks.

As AI services and user experience improve, there is exponential growth in the data volume and complexity. Questflow relies on MongoDB to help them tackle challenges and solve problems. (Learn more about MongoDB for Artificial Intelligence).