Purpose-Built for Research AI is Here - a Letter from the CPO

Today marks a transformational moment for Remesh. Over the past 2 years, our customers have eagerly adopted Generative AI-powered features like Summarize and Auto-Code that have accelerated and deepened their analytical capabilities. But as we looked at the research ecosystem and where technology is heading, it became clear that simply adding more Generative AI tools wasn’t a long term solution. Each new feature demanded training, change management, and extra effort to stitch it into an already fragmented workflow. It was time to transform the research workflow from the ground up. And that is why we launched Remy.

Remy is an AI research assistant that will orchestrate every step of the research process. Our vision is for customers to go from a burning question to a fully executed research plan through analysis, and back to new questions, all within a single agent-powered interface. Today, Remy delivers data-backed assertions as you probe study data with an agility and precision unmatched in the market, based on a set of “tools” that reflect Remesh’s analysis capabilities. 

One of Remy’s superpowers is its ability to evolve its capabilities quickly by bringing online new tools to meet customer demand. We will lean into the strong co-design and co-build partnerships we have developed with customers to identify how we want to upskill Remy next. Today, our users can make requests for new capabilities, and tomorrow, they will be empowered themselves to develop directly into Remy’s agentic architecture. Our eyes are set on a true customizable, end-to-end research workflow. 

‍

How Remy Came to Be

The vision is grand and inspiring, but the origins are humble. In 2024, we saw the potential to solve the last-mile of insights work –- the reporting and presentation of data, with Large Language Models. Our R&D, engineering and product teams focused on reverse engineering exemplary reports so that we could prompt our way to customer-ready reports from raw study data. The result was a systematized way to generate high-quality, in depth reports with a rigor that was commensurate to any gold-standard. But we didn’t ship it. 

Why? Because we realized that in the fast-paced world of quick turn research and same-day insights, our customers were not looking for text-dense hyper-nuanced reporting. Instead, they were looking to answer a set of questions. And importantly, those questions would shift over the course of the study as the team gathered more information: data leads to more questions, which leads to more data, and then more questions once again. We wanted to build a tool that fit naturally with the inquisitive and evolving nature of today’s research, enabling teams to deliver timely and impactful insights — ultimately expanding the role of research in organizational decision making.

Fortunately, our R&D team had already been experimenting with another AI technology that would become critical to solving this problem for our users – RAG, which stands for Retrieval Augmented Generation. RAG is a framework that can be used alongside LLMs to parse large volumes of data - perfect for handling dense qualitative and quantitative data files. The first major step for our teams was to store the data for RAG in a way that was quickly accessible, but also most importantly, secure and partitioned. Trust in the data and trust in the way we handle data has always been a first principle at Remesh. 

From there, we built a custom agentic framework and service to handle non-linear workflows with a tool set that compounds in value, and enables agent to agent communications. It was important to us that Remy was built on a scalable foundation to grow as our client needs change, and the landscape of market research evolves. Remy’s architecture means we can focus on delivering value rather than rebuilding for each innovation in the space. Remy is built to be flexible so that we can accelerate value creation. 

With Remy, the world is kind of your oyster – It allows you to iteratively explore the data in a short period of time creating this sense of momentum and curiosity.

So what can you do with Remy today? Our internal teams have been hands-on for the past few weeks and have been impressed with its range of capabilities from segmentation analysis to study-to-study comparisons and attitudinal clusters.  It doesn’t just save time, it also makes exploring data genuinely enjoyable. Remy creates a flow state that catalyzes curiosity, leading to even more digging and more insights. 

Because Remy is purpose built for researchers, it understands how you can use the data, apply research frameworks and make grounded projections. And, every assertion is grounded in the study with clear citations so you can confirm the analysis and ask clarifying questions, making it defensible by design. One of our consultants reflected: 

“What becomes powerful is this play between data and theory or external research.  I asked it a whole bunch of  strategic questions – If I were designing an ad campaign or if I were designing an employee retention program based on this data, what should I focus on?  Remy surfaced some real small nuanced nuggets, with justification behind the points, that you can only discover as you start to play with the data. It is what I would expect a well-trained, master's level researcher would be able to do .”

We are excited to offer this new format for engaging with qualitative and quantitative data, in a way that complements the existing Remesh ecosystem. Customers who are already well-versed in specific tools and workflows on Remesh may elect to continue to engage in those familiar ways, or innovate on their practice by engaging with Remy. And for others, this new intuitive chat paradigm empowers non-power users to access insights quickly, which is ultimately what we want to make possible. We believe that Remy expands what is possible on Remesh, and who is empowered to understand their products and audiences. If you have curiosity, and an interest in making decisions grounded in data, Remy will get you there. 

It’s only going to get better and more powerful from here. We will keep adding analysis tools, layer in visualizations, and let users export their key findings to slides. What’s more, Remy’s analysis capabilities are just the beginning. Remy is designed to orchestrate the entire research process including study set up and moderation. It’s also designed to work hand in hand with agents that can provide more context or augment Remy’s capabilities. We believe that the future is a workscape that blends humans and agents, expanding the purview and impact of research. With Remy as a powerful agent ally, we want researchers to focus on what they do best: asking bold questions and guiding organizations toward confident, data-informed decisions.

‍

Want to see Remy in action? Sign up for our webinar!