Where chat might be valuable is in inturpreting some of the expressions in language. See the following dialog where I asked it about the frustration level of a customer:
Just to be clear, the AI isnât the compelling part. The first part of Microsoft Fabric is a feature to gather all the data you want to analyze into one place. They use Azure Data Factory to do this and place the results in a Data LakeHouse of files in the same format. It ingests the data from many sources. On the way in, one can apply transformations to the data. This makes it possible to pull in from multiple business systems and then report on data in a common format. It is also possible to fix bad data by mapping part numbers that changed over the years to a common ID for example.
Once in the common format, you now have the ability to run tools like PowerBI over it. Finally, you can build an AI model on just your data. No other company will see this model, and no other data will pollute your corpus. This should lead to less hallucinations and most math will be done outside of the AI with tools that are good at it.
So like @klincecum says, the marketing strategy here is to sell Azure services. To have clean analytical data, it might just be worth it. 
1 Like
I use ChatGPT for debugging Epicor.
The level of confidence between Epicor support and ChatGPT is exactly the same and ChatGPT is right more often (in my experience) than support.
What sort of things have you asked it?
Iâve posted error messages in it and had it help me debug. Iâve asked it for where variables might be exposed. Iâve asked it for help with customizations and BAQs, etc.
One key for that is to make sure you tell it what version of Epicor you are on.
Itâs useful for debugging, looking at logs, and occasionally Iâll toss it some string manipulation requirements and have it spit out regex or a loop because I hate regex. I end up using what it produces less than 20% of the time, but even when it fails it can help visualize the problem. Itâs like writing from scratch vs writing from a prompt. Sometimes you need inspiration.
Itâs helpful to visualize GPT as more of a query engine than something thatâs thinking. The more often it finds an instance of the question youâre asking (followed by the correct answer youâre seeking) the better the output will be. Edge cases, domain questions, and material thatâs behind a paywall or simply isnât frequently found online will often be terrible. It will make up nonsense to fill the gap, because thatâs literally the best fit it can find. It doesnât know that itâs bullshit. It just picks the best fit to complete the pattern in front of it.
Thankfully, Epicor follows most of the conventions of the Entity Framework, and that in turn is a fairly standard CRUD application, built on widely documented .NET and SQL DB technology. So you usually can get decent results. On a good day, it can be like a fresh dev intern who shows flashes of brilliance, but probably should find another major.
Quis custodietGPT ipsos custodesGPT?
2 Likes
Is that kind of like the FBI?
âWe investigated ourselves and determined we did nothing wrong.â
Azure, lake house, fabric⌠soft nice pleasant terms. These names really crack me up. I guess itâs to help programmers fantasize or something. 
I think the Grow demos at Insights are one of the most promising future use cases for AI - having the business be able to ask questions of the data directly, like what was the profit for such and such product or category over such and such time period, would be a tremendous benefit. But I am skeptical the answers will ever be true. In every company I have ever worked for, there is no clear cut way to answer these questions - there are always tons of business rules about if this, except when, and on the 5th Tuesday, that must get built in to various queries to get the real answer. How would AI know any of that? Will we someday be able to train AI with all of these quirks and rules? And how can you validate the result that is being presented when you donât have access to the raw business logic behind it? (or maybe you do have access?).
In its current state, I find chatgpt extremely useful for general time savings like John described - it will read huge long error logs for me and answer questions about them. It can parse xml and find specific values or parameters for me quickly. It will read my MRP logs for me. It will explain errors in my event viewer (although last week, it told me to google the error number lol). It will stub out code for me, It will convert tsql to linq. Tons and tons of time savings for things I have to deal with every day. But as far as hooking it up directly to the ERP - I feel that is still a ways away.
I read somewhere that there is now a marked decrease in traffic on stackoverflow because people are asking their questions to chatgpt instead of eachother. On one hand, its great to eliminate some of the common questions that people seem to ask over and over again. On the other hand, I hope this community (and other tech forums) survive the chatgpt era and we donât stop collaborating just because chatgpt answers faster!
2 Likes
I think once it responds with GIFs weâre in trouble. 
The difference is that ChatGPT doesnât know when itâs wrong and talks out their butt all the time andâŚ. Oh wait⌠we are screwed
How is this any different than you? Are you a robot too!?
And what did ChatGPT use to train on? StackOverflow to some extent. What happens when the training data keeps looking at itself? It might provide some insights occasionally but if the information sources dry up, how good will the models get?
2 Likes
Us. All of us. They scraped our posts, meant to help our peers as a contribution to the community, and are trying to turn it into their property for their profit. Itâs a land grab for our own ideas. The artists out there, to their credit, grasped this immediately the second imagine generators started popping up. Itâs going to take a bit for everyone to realize this is inherent to all generative AI.
And industry is intentionally trying to sleight of hand this by writing vague open letters about how this stuff will lead to Skynet and Terminators murdering the children. Itâs like the plot of a dystopian cyberpunk novel.
The one saving grace is that these programs do not appear to be that hard to fork and iterate. And the most unscrupulous outfits may not be able to commercialize models built on unlicensed training data, since that data is (on some level) embedded within the model and itâs a statistical probability that it will spit out infringing content.
1 Like
one solution to this might be to use an old trick that has been used for decades⌠it is called âSaltingâ the data with known false data that can be traced.
- in the mailing list business, we salted our list with a few random names/addresses that were delivered back to us. These addresses were always sent out with our mailing list ârentalsâ, and they were coded so that we knew the source of the rental. if we saw the address used more than once, we knew someone had broken their agreement to only use our list one time (or they gave/sold it to someone else).
- in the MAP drawing business, there are âPhantom settlementsâ or cities that are inserted onto a map. If that city appears on someone elseâs map, then the original creator of the map knows that someone else stole their copyrighted data (see Phantom settlement - Wikipedia).
Perhaps in the same way, we can insert known salted content that would show up when ChatGPT gives an answer giving us a trace back to the source.
1 Like
A few weeks ago I had it write me a poem about how great Epicor is:
In the world of business, speed is key,
And Epicor ERP is as fast as can be.
From inventory to sales and beyond,
It streamlines processes with a single bond.
Its features are vast, its functions robust,
No task too daunting, no hurdle too tough.
With real-time data at your fingertips,
Decision-making is swift, no time for slips.
Manufacturing flows like a well-oiled machine,
From planning to delivery, every step clean.
Supply chain management is second to none,
With Epicor ERP, everything is done.
Time is money, and Epicor ERP knows,
How to make the most of it, as it goes.
It boosts productivity like never before,
Efficiency and speed, always at the core.
So if you seek a system that wonât let you down,
Epicor ERP is the best around.
With its speed and productivity, youâll surely see,
A world of success, just wait and see.