InfoSol We're passionate about inspired business intelligence
I have noticed an interesting trend over the last 6 months as people and organizations play and try to find productive use cases for AI.
While some companies went all out to encourage their staff to use AI and created leaderboards and rewards for people who burned through the most credits, many approached AI more cautiously and strategically.
One of the most popular use cases has been to create bots or simple user interfaces to just ask AI natural language questions. The questions tend to be analytical in nature like “show me by top 5 performing clients by revenue month to date and the difference from the previous year” and some are more discovery like “how many patients have remained in each ward longer than 5 days?”
As AI code generation capabilities have also dramatically improved with more recent models, you can also ask the AI to create a dashboard or visually pleasing output for you.
So now I see many of the same people who were all out creating Power BI dashboards and reports, just using Copilot instead to do analytics, data discovery and data visualization.
Why not? It’s quicker, easier and, right now, less expensive.
What I am not seeing is organizations replacing operational reporting with AI.
Oh, yes, Gartner said earlier this year that this would be the low-hanging fruit for AI but, once again, they appear to be wrong.
AI is not a good match to replace operational reporting which tends to be tricky due to both complexity and formatting. I have seen some organizations try and get part of the way there but most soon realize they are digging a bottomless pit and throwing good money after bad.
At the same time, many organizations have only recently started their journey with Power BI thinking that it will replace solutions like BusinessObjects. They have already realized it is not simple or low cost to implement Power BI as they usually need a Fabric capacity license along with a team of Power BI experts to implement the infrastructure.
Now, these organizations may want to consider changing course and using Copilot in place of Power BI. After all, Power BI is a tool created for analytics, data visualizations and data discovery – all of which Copilot and other AI tools can do.
This also brings up the question of dashboards which appear to be on the decline. With the overwhelming amount of information being thrown at all of us today, we are looking for smaller and more concise summaries. Rather than a dashboard, that we must spend time interacting with, we just want the key details or a concise summary. AI is being used more and more for those concise summaries. In addition, more people prefer to receive a text of the key information they need rather than a dashboard or report.
A recent addition to InfoBurst, the platform for information scheduling, bursting, grouping and distribution, allows delivery as an SMS text message.
To anyone looking to convert operational reports to AI, I would recommend re-thinking that strategy.
However, AI replacing Power BI for analytics and data discovery is a real possibility and something worth considering.
Hi Paul,
Excellent info and a little relief for me who has been developing SAP BO universes and WebI reports which can sometimes be very formatted. In power BI they call it paginated reports.
I am learning Power BI and plan to take the PL300 exam and I also plan to take Fabric Cert.
I do need your advice. For someone who is BO developer, what must he become good at or learn to stay valid in the age of AI. I think it could be semantic modeling or Ontology that can be utilized by AI chat box like CoPilot.
Thanks for the positive feedback. I feel that BusinessObjects reporting will remain relevant and in need for at least the next decade partly because it is embedded in so many applications and partly because there is no next generation reporting solution on the market.
As a BO Developer, yes, it would be good to master IDT and the semantic model which will be helpful with AI going forward. Also stay up to date with the latest Webi innovations – there’s a lot coming in BI 2027 next March.
Given the recent revelations about AI by the people working at those companies building and deploying these applications I’m not certain that I can give credence to the results of any AI engine no matter how complete it seems to be> Apparently, those who know have indicated that AI may have already gotten smart enough to tells us incomplete facts based on the bias of the engine. Hopefully, it will result in many professionals who have been released from jobs because of AI to be brought back into the workplace to resume practicing their craft and producing credible, reliable and verifiable results based on learned skills.
AI needs responsible management. It is a great assistant but it is not always correct.