ChatGPT enables business leaders to get answers to business questions by asking a question and getting an answer. So, will ChatGPT change the way we consume business intelligence (BI) derived from data as well and therefore our use of BI tools like Power BI, Tableau & others?
Today, business intelligence is often consumed by viewing dashboards – you might call them trackers, spreadsheets, logs, or similar. But usually, business data is presented to business leaders in some kind of structured format, whether in numbers or graphs. Think about…
- Board reports on financials
- Marketing campaign dashboards
- Call centre performance statistics
- Customer satisfaction & complaints
- And so on
We all expect to see business data presented to us in the same way every week or month, and we get used to the format and layout and the way that metrics have been calculated in the same way every time.
We also know that we if we ask someone a data-related question, unless it’s already included in one of these dashboards, someone will have to go get the data and run some statistics in a spreadsheet, in order to answer the question…which could take hours, days, or even weeks. We’ve all been there, haven’t we?!
But imagine a business plugging a tool like ChatGPT into all their data systems…in future business leaders could simply type, or ask a question, and get an immediate question. Think of questions like…
- Are we under or over-target on revenue this week?
- Which region performed best last month?
- Give me a summary of our KPIs year-to-date
- Tell me which metrics I need to worry about
- And so on
So, with such easy access to business intelligence, will business leaders need dashboards at all in the future? If the answers to all business intelligence questions are on the end of a question, why bother to check a dashboard? Just ask a business question and get a data-powered answer. Easy!
But, if this was the case, why hasn’t this happened already? Microsoft Power BI has had a natural language processing (NLP) capability, where you can ask it to show data in any format of your choice, live for 7+ years and yet it’s rarely used.
So, are humans truly curious and do we really like to explore anything and everything we like, all the time? Or are we lazy, habitual, and just want to be served business insight on a plate instead?
Of course, it’s a mix of both…
- ChatGPT-type functionality will help businesses transform faster to a self-serve data-informed organization which empowers people to be curious and ask any question they like about anything, at any time, in any way they like
- whereas dashboards will ensure that data is presented to the right people when they need it, in a predictable way, so they don’t forget to ask the most important questions
So, dashboards are dead…long live the dashboards…?!
Check out this article on Forbes – https://www.forbes.com/sites/forrester/2023/04/28/the-current-state-of-llms-in-business-intelligence-and-what-needs-to-change/?sh=4e1d433147d8 – as these 2 key points stand out:
- Use the latest tech to source data – businesses could massively benefit from using third-party data to enrich their own internal business data as the combined data would be much richer and they could compare their business performance to the wider market
- Use guided analytics – businesses could use data and chat-like capability to suggest what to ask of your data next such as “here is the next step in the analysis” or “since you asked that question, you should also ask the following questions” in a ‘next-best-action’ way
So, it all sounds amazing right? Let’s go! There are so many opportunities that will enable business leaders to interact with data in easier, quicker, and more valuable ways using the latest technology.
But there are two fundamental problems we see when using NLP for data analysis:
- Mapping to user needs – pinning down specific use cases is hard. Really hard. NLP would be awesome in Board or strategy sessions to answer a broad range of questions on-the-fly. But people like to be prepared for meetings, so the need for on-the-fly analysis is just one part of the solution, not the entire solution.
- Consistency – what most people want is consistency. We all like to know what’s coming don’t we? So, we need a familiar set of KPIs that we can track movement against and set targets for. NLP struggles with consistency as it’s not designed to be fixed which makes it hard to hold onto something specific and action insight off the back of it. It’s also hard to predict. Sometimes you ask for something and the answer is astounding. Sometimes it’s nonsense.
So, for most businesses, we see a world with two types of data solution:
- Dashboards – whether you call them reports, trackers, logs or dashboards, we think business leaders will always want information presented to them in a consistent way
- Chat-like data interrogation – regardless of the technology deployed, we think business leaders will love the on-the-fly approach to data interrogation so they can get answers to any business questions on demand
How much does it cost to get up and running with NLP? You may be thinking you don’t have the time or the budget to start using NLP and chat-type functionality. You are wrong! Power BI has native NLP capabilities, for one. This means we’re talking hours, not days, to add NLP functionality to your Power BI dashboard.
As with all new toys and new technology, the way we effectively learn is to play with them. NLP can fill those minutes before meetings and on trains. You’ll be amazed at how quickly you’ll pick it up once it’s at your fingertips. So, have a play or get in touch and ask us about adding NLP to your dashboard today.