Case Studies

How AI helps Marvin's users spend 60% less time analyzing research data

Learn how Marvin's product team embedded powerful AI models into the user research platform to help end users spend less time wrestling with data and more time making sense of it.

How AI helps Marvin's users spend 60% less time analyzing research data

Companies need trained researchers to dig deep and understand customers’ biggest pain points to build customer-centric products that compete in today’s hypercompetitive markets. Thankfully, significant strides in AI research—like the research behind Stable Diffusion, modern Large Language Models, and Poisson Flow Generative Models—have now made AI a formidable co-pilot to help companies automate this process and extract the most valuable insights.

Marvin, for example, is a qualitative data analysis platform building powerful features and tools on top of the latest advances in AI, like AssemblyAI’s state-of-the-art speech recognition model Conformer-2.

Marvin helps its end users collect, organize, analyze, and share qualitative research data in order to guide them to ask the right questions, make sense of patterns, and build better customer-centric products and services.

To accomplish this, Marvin’s product team relies on a variety of technological tools, including AI.

The Marvin product team has embedded powerful AI models like AssemblyAI’s Core Transcription and PII Redaction models into the user research platform. With these advanced AI models, Marvin’s users spend less time wrestling with data and more time making sense of it, driving improved ROI and intelligent insights.

In fact, users spend 60% less time on average analyzing data when they have a centralized platform such as Marvin, which frees them up to focus on other priorities.

Source: Marvin

Knowing users choose Marvin to process and analyze gigabytes of audio and video data, Marvin’s product team needed to build highly accurate AI transcriptions to serve as a standalone feature and a foundational model that feeds into Marvin’s qualitative data analysis features.

Marvin’s product team works with AssemblyAI, knowing they can provide both.

“AssemblyAI is doing amazing work, and we rely on AssemblyAI to be the tool of choice, especially when it comes to audio and video intelligence and transcription,” explains Marvin’s CEO and Co-Founder Prayag Narula.

Marvin’s product team is excited by AssemblyAI’s commitment to building an ongoing partnership, as well as the company’s proven ability to ship secure, scalable, and robust AI models, which allows them to deliver a best-in-class service to their end users.

For example, AssemblyAI’s new Conformer-2 model, a state-of-the-art speech recognition model trained on 1M hours of English audio data, achieves near human-level performance across a wide range of data. This means the audio and video data processed by Marvin’s platform will be even more accurate, and more valuable for its end users.

Learn more about Conformer-2

With hundreds of hours of research data, interviews, and confidential information being processed, data security and privacy are a top priority for Marvin and its customers.

Marvin needed an easy solution to help users strip Personally Identifiable Information (PII), such as phone numbers, addresses, and company financial information, from the audio and video files they processed.

AssemblyAI’s PII Redaction model enabled the Marvin product team to build automatic privacy filters that automatically remove PII data. This capability was the first of its kind in the user research industry and ensures all of the data users processed stays secure.

With AssemblyAI, Marvin also released Live Notes with advanced AI, AI summaries, automated tagging and note-taking, and other tools that integrate with Zoom, Google Meet, and other video conferencing platforms. Instead of taking manual notes during user interviews or internal meetings, Marvin’s customers use AI features to accurately automate note-taking and summarize key conversational areas.

Source: Marvin

Marvin’s leadership and product teams recognize the significant impacts AI and Large Language Models have on the tech industry. They take the responsibility of being an early AI adopter very seriously.

“When it comes to something as transformational as AI, as Large Language Models, I think it's our responsibility as technology builders to really understand and predict and study the impact that this technology can have,” says Narula.

Marvin is committed to sharing this responsibility with partners like AssemblyAI who take a proactive, transparent, and thoughtful approach to building AI technologies, and to protecting and empowering its users during this AI transition.

Want to learn more about building AI-powered tools? Watch key moments from the interview, read the transcript, and more on the Marvin platform.