Crunch AI
Crunch is an Al powered data analytics tool that helps you uncover the deepest insights from your data
Overview
Crunch is an Al powered data analytics tool that helps you uncover the deepest insights from your data
Product teams rarely suffer from a shortage of data. They have events, funnels, retention charts and dashboards full of numbers. The problem is that reaching a useful answer still requires knowing what to track, how to structure the analysis and which question to ask next. Crunch was designed to shorten that journey. It automatically captures activity from a connected product, helps teams organise those events and lets them explore their data by asking questions in plain language. Instead of stopping at a chart, users can keep digging into individual data points, follow suggested lines of enquiry and build a visual trail towards the cause of a change. The aim was not to build another analytics dashboard with an AI chat box attached to it. It was to redesign the workflow around the way product teams actually investigate a problem: one question leading to another.
Challenge
Analytics tools asked for too much work before they became useful.
Setting up product analytics can take weeks. Engineering teams need to install tracking, define events and make sure everything is named consistently before anyone can begin learning from the data. Even after that work is complete, users are often left with a different problem: a wall of dashboards. A chart might show that conversion dropped or retention changed, but it rarely explains why. Finding the cause means creating more reports, changing filters, comparing cohorts and remembering which thread of the investigation led where. Crunch needed to make both ends of that experience easier. Setup had to feel manageable for a small team without a dedicated data function. Exploration had to work for someone who understood the product problem but did not necessarily know how to write a query. And the AI needed to help without turning the experience into an open-ended chatbot that produced confident but contextless answers. The challenge was to make powerful analysis feel approachable without flattening it into something shallow.
Approach
Reduce the distance between noticing something and understanding it.
The design followed the full analytics loop rather than focusing only on visualising data. A team first had to connect its product and confirm that the right events were arriving. Those events then needed to be labelled in language the wider team could understand. Once the data was ready, users needed a simple way to begin an investigation, explore possible causes and continue monitoring anything important they discovered. Every part of the experience was designed to make the next action obvious. During setup, that meant clear installation steps, a live event feed and assistance with event naming. During analysis, it meant suggested follow-up questions, contextual AI and a canvas that preserved the path of the investigation. Afterwards, it meant turning an insight into an alert without moving into a separate tool. The goal was a continuous workflow: connect, understand, investigate and monitor.
Research
A chart usually tells you what changed. Perhaps fewer users completed onboarding. A particular cohort retained better. One version of a feature was used more than another. The useful work begins immediately afterwards: Why did it change? Which users were affected? What happened before that event? Is the pattern isolated, or does it appear elsewhere? Traditional analytics tools treat each of those as another report to build. Crunch instead treats them as branches of the same investigation. That insight became the foundation of the product. Analysis would not live inside a collection of disconnected dashboards. It would take place on a canvas where every follow-up question retained the context of the one before it.
Competitor Analysis
Many competing icons depended on gradients or complex illustration that became muddy at small sizes. The opportunity was to create a cleaner mark with a stronger silhouette and more controlled depth.
Visual Direction
The visual direction uses bold form, precise curves, and a focused sense of dimensionality. The icon feels polished but not over-rendered, with details chosen for legibility first.
Design solutions
Turning a small canvas into a memorable brand asset.
The final icon system balances depth, simplicity, and product personality. It was designed to work across the app store, device homescreens, marketing surfaces, and dark or light contexts.
Typography
Typography was treated as supporting material around the icon rather than the main identity device. Labels and supporting copy stay neutral so the mark remains the focal point.
Layout System
Presentation layouts emphasize scale testing, contrast comparisons, and side-by-side usage examples to show how the icon performs in realistic environments.
Components
Supporting components keep the case study structured: preview tiles, comparison rows, and usage frames make the process easy to scan without distracting from the icon itself.
Outcome
A compact visual mark with product-level presence.
The final icon feels recognizable, scalable, and ready for real product use. It gives the app a stronger first impression while preserving clarity across every size.
