From Queries to Conversations: Simplifying the Last Mile of Data Access

Data analytical tools are complex. With a prerequisite of learning to use data tools to begin; the tools create an entry barrier for non-technical users. Once you have this skill, as a data analyst, you spend weeks effort, building dashboard by going through individual data column before you can even glimpse your your data

Now, even if you built the most perfect dashboard, they fail because executives lack the time to interpret charts, and business users avoid cognitively overwhelming interfaces, relying on analysts for weekly digging, pattern matching, and storytelling. This reliance persists because current tools don't invite or enable easy data exploration.

Most data visualisation tools today are built like Photoshop, very powerful, but demanding tool expertise and a lot of cognitive effort in comprehending the data. What if we designed these products differently to solve the last mile of data access? What would be the product principles required to create the BI equivalent of Canva and how does it translate to UX?

Today, AI should handle mundane tasks. For creators, this means automating artifact creation and repeatable data analysis. For consumers, it means delivering insights in engaging formats to bridge the interpretation gap.

However, integrating generative, probabilistic AI into a complex, accuracy-dependent tool is challenging. Its unpredictability makes it less trustworthy. Creators need deterministic outputs, and enterprise users demand trusted data for critical decisions, yet also need engaging data interaction. How can we balance automation and human control? Can existing simplicity principles suffice, or are new guidelines necessary?

Outline Business dashboards are often boring and hard to use, making insights difficult to grasp. We will start by engaging the audience with a problem identification exercise to tailor the talk and segue into how current BI tools fail both consumers and data analysts.

We will then establish the product principles used in existing BI tools and share our process for developing unique principles that yield a different user experience, showcased through UI examples and system integration.

Next, we will challenge these principles with emerging AI technology. We'll set the context for AI's benefits in business analytics, discuss industry challenges, and examine which product principles remain valid and which need amendment in the age of AI.

We will conclude by addressing the open-ended design questions the industry faces regarding a rapidly evolving AI landscape.

Audience Designers, product thinkers and visualisation practitioners,

Takeaways This talk offers actionable workflows and a framework to simplify data visualization for practitioners and broaden product accessibility. For designers and product thinkers, it provides a practical framework to simplify their own product, avoiding deep design philosophy.

Speakers Kshipra Sharma, Design leader for ThoughtSpot's Connected Insights Track, her work focuses on user experiences in charting, dashboard building, and leveraging AI for direct consumer insights.

Faris Mohammed is a Lead Product Designer at ThoughtSpot, shaping the charts and search experience to make it simpler and more approachable, especially as AI reshapes the industry.