Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

ChatGPT Is Much More Than A Data Prep Assistant

ChatGPT isn’t merely an AI—it’s humanity’s collective knowledge and has profound implications for how organizations will access and use data

The business intelligence (BI) and analytics industry is already using OpenAI’s ChapGPT to automate tedious data preparation tasks. ChatGPT (and other generative AIs) even allow organizations to ask questions about their own data using natural language. Though impressive, these use cases don’t even come close to capturing ChatGPT’s full potential. 

ChatGPT isn’t just a data labeler, query writer, and coder. Rather, ChatGPT is a pretty accurate database of humanity’s accumulated knowledge (except when it hallucinates), accessible through a Q&A interface. That has profound implications for how and when organizations will pursue answers to difficult questions. 

The Questions We Don’t Answer Today

Let’s say you run an ecommerce website and see a strange, two-day dip in sales volume in a specific location.


Top AI Machine Learning News: Instabase Doubles Valuation to $2B and Launches AI Hub, A Generative AI Based Platform for Content Understanding

Today, you would need to generate hypotheses and then look for databases to test them. Was the dip caused by weather? Labor strikes? A local competitor’s sale? A news story on layoffs or financial turbulence? Testing these hypotheses one by one is so inefficient that no organization would bother to do it, even if the answer could be valuable. 

Once humanity’s collective database, ChatGPT, is up to date and not stuck in 2021, you’ll be able to ask it for all the key events that happened at a certain location on a certain date. In minutes, you’ll have a data-backed theory of why sales dipped. Thus, your organization will gain the power to investigate events and decisions where time, location, nature, culture, and countless unknown variables previously made an inquiry infeasible. 

How BI and ChatGPT Work Together

If, eventually, everyone can ask ChatGPT questions about data using natural language, why do we even need BI and analytics systems? 

Related Posts
1 of 6,035

BI is a repository for internal, proprietary data and a tool for making that data available, useful, and actionable in applications. BI also allows you to map your organization’s specific data into knowledge (e.g., what is “revenue” and what defines a “department”?). ChatGPT is for asking questions about things happening outside your organization. It taps into datasets that you wouldn’t necessarily check, finding correlations that illuminate the meaning of your proprietary data.

Every BI and analytics system will need to tap into ChatGPT. But no, ChatGPT won’t replace them.  

The Limitations of Humanity’s Database

ChatGPT has three weaknesses relevant to our discussion. The first is accuracy. While GPT-4, the newest language model from OpenAI, is more likely to give factual answers than its predecessors, it still makes mistakes. For most organizations, that’s ok. Speed is more important than perfect accuracy. 

The second issue is security. Organizations are nervous about sending their data to a central AI engine that retains the information for training purposes. Thankfully, new versions of ChatGPT will not keep the inputted data of enterprise customers. Problem solved. Just as companies grew accustomed to external cloud platforms like Salesforce, they’ll get used to external AIs.   

The third problem is attribution. Often, ChatGPT surfaces information but cannot direct you to the source. Attribution will take time to fix, and again, that’s ok for most businesses. While academic researchers and journalists (should) have stringent standards for attribution, businesses typically don’t. If vetting a source is essential, someone can track it down the old fashioned way with a search engine. That will still be much faster than the alternatives.  

Analytics Without ChatGPT Will be Unthinkable

ChatGPT is not merely an efficiency and automation play. It and other large language models will represent humanity’s accumulated knowledge in a highly accessible form. It might be a questionable mathematician, lousy stock picker, and uninspiring screenwriter (for now), but that says more about our expectations than ChatGPT’s capabilities. We don’t need ChatGPT to replace artists—we need it to better understand our organizations and the complex world they inhabit. 

Any technologies that include or embed analytics must be built to harness humanity’s accumulated knowledge.

If not, they’ll become irrelevant.

Recommended: Top 10 Gen AI Platforms that are Driving Transformation In B2B

[To share your insights with us, please write to]

Comments are closed.