During the design and creation process of our Analytics product, we spent a lot of time thinking about the best way to show the data and the insights to be analyzed. Dashboards are very useful as data discovery and exploration tools. At Lexic we work with unstructured data that is often not exploited in companies and are difficult to show in dashboards.
That is why we understood that a dashboard was crucial to understanding this "new data".
But were we right?
The reality is that it is, but only half. The discovery of the data is very important. Becoming familiar with new data that can add value is crucial to be able to make it actionable. For this, you need a dashboard. That first phase where we discover a new dataset is very important and the use of the dashboard is very intensive, but what happens next ...?
Having understood this, how can we add value to the process of data discovery and go beyond classical analytics?
Companies already have many dashboards: ERPs, Business Intelligence, Web / App Analytics, etc. Do we need a new dashboard 😅?
The use of AI allows us to design a new product strategy to make data discovery more intelligent and efficient. We are currently working on really interesting concepts from a product perspective:
1- Dashboardless: as the name suggests, the idea is that the insight and KPI to be measured reaches the right person when needed. Companies do not have data problems but time and resources to "heal" it, to filter it. Our customers are testing that concept through the "Notebooks" and "Alerts" functionality. These functionalities allow the system to intelligently decide (hybrid approach: the system decides because an expert has set a series of rules and settings) when to report on a certain event (something grows or decreases), data, or insight.
2- Gaming vs Dashboard: "let the client play with their data". We repeat this phrase constantly because it works. We have prioritized elements and tools that allow the client to play with the data, cluster-shuffling, auto-segmentation, auto-labeling, semantic search engines, personalization of their own "relevant concepts", etc. We have turned the visual dashboard into a dynamic and exploratory dashboard.
3- Prediction: predictability is a classic element in AI but it is less so when we talk about predicting the behavior of, for example, customer conversations during their interactions with customer service centers. At Lexic we have an internal project that tries to replicate precisely this. The objective is for the No-Dashboard to be able to warn us that something "important" (error in payments, complaints about a new product, ...) is about to become important.
We envision in Lexic a content curation, insights, and alerts that work in an intelligent and automatic way where the role of the dashboard is limited to the AI ​​receiving the input (human help) of the business expert, of a person. Insights are triggered and delivered to your mobile device, to your Slack, to your Teams. We finally imagined a chatbot interface that allows us to ask the AI ​​"how are the delivery "delays" going on today?" The Bot would be able to curate the most relevant information and display it for you.
We are so close. Another step towards our purpose of putting the person at the center of AI.
Comments