We’re seeing a wave of new automation use cases made possible by Grooper’s integration with Large Language Models (LLMs). These aren’t theoretical; they’re solving real problems that just a few years ago couldn’t be automated.
Let’s look at three examples, ending with one that’s already live in production…
#1: Expense Report Automation
Expense reports have always been a pain for employees and accounting teams alike. That’s because expense policies vary from company to company, with exceptions like:
- No alcohol
- Per diems
- Regional hotel limits
- And the ever-tricky “how much tip is too much?”
Traditionally, enforcing those policies meant accountants combing through receipts. But now Grooper can automatically flag violations, like detecting alcohol purchases (even in other languages). Staff review only the exceptions, not every receipt.
The result? Faster approvals, happier accountants.
#2: Sales Order Automation
Sales order processing sounds simple until you realize how many customers, SKUs, and item descriptions are involved. One wrong part number can trigger hours of manual research.
Grooper can combine LLMs and vector-indexed data to semantically match product descriptions. Think “Google search for part numbers” but on steroids. Grooper compares what was ordered to what’s actually in inventory, flagging mismatches instantly.
What once required a lot of manual work from sales admins can now be done in moments.
#3: Price List Automation (Client Onboarding)
Some businesses onboard clients by loading entire price lists into their systems. It’s a time-consuming, error-prone, manual process.
Grooper’s AI Extract feature now lets you automatically normalize inconsistent price lists into a standard format. Our friends at Funeral Directors Life are already doing this in production today. They’re turning what was once 100% manual data entry into a quick, repeatable process, saving hundreds of man-hours annually.
If you’ve ever said, “We tried automating that before, it just wasn’t possible,” it’s time to take another look. Grooper’s AI capabilities are growing fast, and what used to be impossible is quickly becoming routine.
Local AI?
I recently wrote about our big Cray AI server, which we’ve nicknamed “Aristotle.” It’s a beast, and it costs a significant amount for us to do local AI development with.
But this week, NVIDIA’s GDX Spark was made available. It’s a small, local supercomputer with 128 GB of RAM and enough juice to run huge open-source models (up to 200 billion parameters).
The age of local inference is here. That’s why we’re testing our AI server – so we can recommend options beyond cloud offerings to you.
Many of our clients want to benefit from AI but can’t send their sensitive data to the cloud. Local LLM AI Servers solve that issue.
There are affordable, scalable local solutions available in every price range now.
So reach out to us if you are interested in how a locally installed and managed LLM can change your company!
New Blog Alert!
We have a new blog posted that came out of my research around accounting in higher education.
As it turns out, accounting in higher education is more complex than typical corporate accounting because it involves external funding sources such as grants and donations.
While the blog talks about FOAPAL and how one specific system (Ellucian) handles this process, the same applies to any chart-of-accounts-type accounting system.
Who knew? I bet you accounting gurus did!
