Scoops

An AI-powered research companion

Scoops is an AI-powered qualitative research tool. It enables marketers and brand managers to generate a qualitative survey, send it their customers, and analyse the data in minutes. Our tool applied the power of Large Language Models to accelerate the research process that typically takes weeks.

As a small team we designed, built and launched one of the first AI-powered research tools. We learned a ton about how to engineer prompts and deploy LLMs in working products. We developed a deep understanding of leading models (ChatGPT and Claude), using different models for different tasks depending on their strengths.

The branding and experience design was intentionally approachable—designed as the antithesis to the swathe of cold, purple AI products that emerged during 2023/24.

Our initial concept sketch was used during early research and helped define both the concept and our ideal customer profile.

Our initial concept sketch was used during early research and helped define both the concept and our ideal customer profile.

Our initial concept sketch was used during early research and helped define both the concept and our ideal customer profile.

The branding and UI captures a playful energy, designed to make the daunting task of running research feel more fun and approachable.

The branding and UI captures a playful energy, designed to make the daunting task of running research feel more fun and approachable.

The branding and UI captures a playful energy, designed to make the daunting task of running research feel more fun and approachable.

A radically simple three-step process enables us to feed our LLM with the context necessary to generate nuanced, qualitative questions.

A radically simple three-step process enables us to feed our LLM with the context necessary to generate nuanced, qualitative questions.

A radically simple three-step process enables us to feed our LLM with the context necessary to generate nuanced, qualitative questions.

Respondents complete surveys in a chat interface. Our AI emulates a human researcher, gathering depth that is difficult to attain through traditional form-based surveys.

Respondents complete surveys in a chat interface. Our AI emulates a human researcher, gathering depth that is difficult to attain through traditional form-based surveys.

Respondents complete surveys in a chat interface. Our AI emulates a human researcher, gathering depth that is difficult to attain through traditional form-based surveys.

Collected data is automatically analysed through a chain of prompts that give the LLM an opportunity to pass through the data mulitple times to formulate insights from observed patterns.

Collected data is automatically analysed through a chain of prompts that give the LLM an opportunity to pass through the data mulitple times to formulate insights from observed patterns.

Collected data is automatically analysed through a chain of prompts that give the LLM an opportunity to pass through the data mulitple times to formulate insights from observed patterns.