top of page



Generative BI – The Next Evolution in Self-Service Business Intelligence

Brewster Knowlton, Founder & CEO

Gemineye

Read More

A Crawl, Walk, Run Approach to AI Adoption

Apiture

The AI Revolution: It's Time for Credit Unions to Lead the Charge

CU NextGen

Using AI to Drive Outcomes and Efficiencies for Community Financial Institutions

Cotribute

How AI is Reshaping Credit Union Operations and Member Experience

Eltropy

Revolutionizing Banking: How AI Agents Will Redefine Personalization and Transform Your Financial Decisions

Finalytics.ai

Generative BI – The Next Evolution in Self-Service Business Intelligence

Gemineye

Overcoming Compliance and Data Challenges in Financial Services with AI

Hapax

Innovation and Inclusion in AI Credit Risk Models

Modelshop

Collaborative Intelligence: Humans and AI Joining Forces in the Financial Sector

Reseda Group

Beyond Credit Scores: AI for an Inclusive Financial Future

Scienaptic AI

Tyfone’s Srikanth Says AI Is Leveling Playing Field Between Haves, Have-Nots

Tyfone

We’ve all become enamored with LLMs since late 2022 when ChatGPT was released to the public.  Countless articles have been written about the current and future use cases for LLMs and how they could affect the way businesses operate.  Yet most of you reading this would likely agree that you haven’t found a way to fundamentally improve your credit union operations with LLMs. 

 

Have you done some fun things with it? Yes.

 

Has it saved you some time scouring StackOverflow for coding help? Yes.

 

Has it fundamentally changed an aspect of your credit union operating model or key processes?  I’d venture to guess the answer is no.

 

The hard part about LLMs is they often lack the context and security needed to provide answers relevant and specific to your credit union.  They are, inherently, large.  Large is great when looking for generality – small and niche is where we need to be focused for our specific credit unions.  In industries with sensitive data environments – like credit unions – it can be especially difficult to efficiently train an LLM with tools like ChatGPT given the security and PII concerns.

 

If you’ve ever thought about asking ChatGPT “how many new deposits did we gain last month” or “how many indirect auto loans are on our books today”, you know ChatGPT won’t have the answers you’re looking for.  Databricks’ new feature, Generative AI/BI, could be the game changer you’ve been waiting for. 

 

Through a fine-trained LLM that privately and securely uses your business intelligence team’s metadata, example queries, and a feedback mechanism based on user’s satisfaction with the response, Generative BI democratizes the way staff can ask questions of their data.  The days of having to send in a ticket to the IT or data team asking “can I get a report of all credit union members over the age of 18 with less than $20 in their checking account and an open indirect auto loan that was opened in the last 30 days” can become a thing of the past. 

 

Using a ChatGPT-like natural language interface, the future of self-service business intelligence may rest on interactions like this: 



 Termed “Generative BI”, this next iteration of data intelligence overlaps the worlds of conversational AI and business intelligence.  Imagine being in a Board or Executive meeting being able to ask, in plain English, questions about the credit union – from retail operations to lending to marketing to finance - and getting validated answers without having to write SQL queries from scratch or navigate Power BI dashboards mid-meeting.

 

At Gemineye, we always talk about the “question-answer feedback loop”.  At your credit union today, if you ask a question of your analyst or data team, you’ll likely get a spreadsheet or visualization back sometime between a day and two weeks from now.  Inevitably, that report or visualization will answer a question (or a part of a question you have) but likely lead to more questions as you dive into the data more.  This leads to another question – and potentially another week-long wait to hear back from the data team or analyst.  After a few iterations of this question-answer loop, six weeks could have gone by from the initial question…and you’ve completely forgotten about why you were researching it in the first place.

 

With Generative BI, a non-technical individual can shrink that “question-answer feedback loop” to near zero by asking questions through a simple chat interface, get an answer back, and then continuing to dive into the data further. 

 

What could you do if you had a ChatGPT based on your credit union’s data?  Fortunately, you no longer have to wonder.

Rectangle 104.png
Group 1000001799.png

Request More Information From This Provider.

Group 16.png

©2024, Finopotamus LLC. Copyrights for all articles belong to their respective authors; used by permission.

bottom of page