Use the Data Insights Agent

Hi, I’m Taylor Baker, Product Manager for the Data Insights Agent in Adobe Customer Journey Analytics. In this quick demo, I’ll show you how the Data Insights Agent makes data analysis easier for everyone, not just data experts. We built the Data Insights Agent because our customers were telling us that their team struggled to access insights on their own. With this tool, now anyone can ask a question like, show me leads from Marketing Campaign X and get a clear visual answer instantly. Let’s jump in with a demo. Today, I’ll be playing the role of a user who’s trying to explore some of my data for the first time. Previously, I’ve relied on Data Analyst in my company to answer my questions, but with the release of the Data Insights Agent, I can now build analyses and answer those questions on my own. First, I’ll open the AI Assistant chat rail by clicking the chat button in the shell header, and then I’ll review the welcome message that gives me an overview of the capabilities of the Data Insights Agent in CJA, what it can and what it cannot support.

I’m interested in understanding the purchases that my business received earlier this year, so I’ll start by asking, compare purchases by product category from March to April. The Data Insights Agent looks through all of the data in the data view, including the different types of metrics and components, and translates my prompt to pick out the right dimension, metric, and date range for this analysis. As you can see, it not only generated a freeform table that shows the correct time range, metric, and dimension, but it also selected the best visualization to showcase the data and answer my original prompt. Here I picked a bar chart to easily compare two different time ranges, and then added a percent change column for additional insight consumption. Let’s say since I can see that the hunting category is the most profitable that I want to dig in there. I could follow it up with a prompt like, proportion of profit by subcategory so far this year for the hunting category. The Data Insights Agent will look through the data view, and again find us the best match of data, the best visualization to showcase this, and filter down the panel for just the hunting category. So now these subcategories are filtered down to only those that are part of the broader hunting category. And because the Data Insights Agent is generating workspace visualizations, I can still interact with visualizations and access all the core features of workspace that users love, like intelligent captions. So all you can do is just click on this intelligent captions button here, and I quickly have insights for that viz.

The Data Insights Agent can therefore help users onboard into workspace by enabling them to discover these features over time instead of being dependent on them from the beginning. Finally, now that I know I can see the game cameras are our most profitable subcategory, I can trend that out and start to dig in a little bit more by just saying, trend profit for the game camera subcategory so far this year. The agent quickly will build me a line visualization, filtering the panel for the game camera subcategory. I’m excited I’ve been able to answer my own data questions instead of relying on a data analyst at my company, and now I have a project of insights I can share with my manager. I’ll wrap up by highlighting a couple of the other key benefits of Data Insights Agent.

First update functionality. In addition to generating new visualizations, Data Insights Agent can also update previously created visualizations. So for example, if I asked, show me revenue by product category last month, but I also wanted to see orders data, I could simply follow it up by typing add orders.

And my prior chart would then be updated to add in orders data. So now I have both revenue and orders data in this bar chart. Second, clarifying questions. If a user asks an ambiguous question or requests a component that doesn’t exist, like show Japan revenue last month, where Japan doesn’t exist, clarifying questions will help return a message suggesting alternative prompts, helping users find the component they were looking for. So in this case, let’s say I was looking for regional revenue last month, I can simply run that, and I then have regional revenue last month. Data Insights Agent represents a significant leap forward in self-serve data analysis in CJ. By expanding access to new personas, we’re empowering more users to quickly unlock insights than ever before. That wraps up the overview of the Data Insights Agent in CJ. Thank you for your time and interest.