JD EDWARDS SCENARIOS
We Want Employees to Ask Questions About JD Edwards Data in Microsoft Teams, a Web Portal or a Chatbot. What Solution Should We Use?
Teams is where the answer appears, and the project is making that answer correct, permitted and traceable.
- Delivery surface and answer quality are separate problems. Getting the surface right and the substrate wrong produces a fast, convenient way to distribute wrong numbers.
- Permissions become more visible when access widens, because people who never had a JD Edwards login start asking questions.
- The session must carry the requesting person's role, or you are building a second security model to maintain and trust.
- Expect different question types by audience. Operations asks about orders, finance asks about balances, and the same substrate serves both.
- ChatFin delivers into the surfaces your people already use while queries run inside authenticated sessions with the rows attached.
This request usually comes from a sensible place: the data is useful, most people cannot get at it, and everybody already lives in Teams. Widening access is a genuine productivity opportunity.
It also changes your risk profile, because you are about to let people ask questions who never had a route to the ledger before.
The Surface Is Not the Project
Connecting an answer to a chat surface is a well trodden integration. What determines whether the deployment succeeds is everything underneath it, and that layer is the same whichever surface you pick.
This matters because it means you can choose the surface on convenience grounds and should choose the platform on substance grounds. Getting that the wrong way round produces a very convenient way to distribute unreliable numbers to a wider audience than before.
Asking the ERP a question
Putting JD Edwards inside Microsoft Teams makes asking easy. Whether the answer is right depends on the work underneath, which is the actual project.
I can almost chat with my ERP system. I can chat with my P&L, chat with my balance sheet, and understand what is causing this flux.
Jason Yocum, on the AgentCFO podcast
Spend the effort on grounding and permissions. The chat window is the part users see and the smallest part to build.
Permissions Get Harder When Access Widens
A conversational surface inside Teams reaches people who never had a JD Edwards login, which is the point and also the risk.
The clean answer is that queries run inside an authenticated session carrying the requesting person's environment and role, so entitlement is inherited rather than invented. The alternative, a shared identity with filtering applied afterwards by the chat platform, means maintaining a second security model and hoping it agrees with the first.
The Answer Quality Requirements Do Not Relax
If anything they tighten, because a conversational answer in Teams carries an air of authority and gets forwarded. A wrong number in a spreadsheet is one person's mistake; a wrong number pasted into a channel becomes organisational fact within a day.
Different Audiences Ask Different Things
Widening access changes the question mix, and it is worth anticipating. A single substrate serves all of these, but the framing and the permitted scope differ.
In practice the operational questions arrive in far higher volume than the finance ones, and they are frequently the ones that save the most time because they currently interrupt somebody.
What to Log and Why
Logging in this context serves two purposes and the second is the one nobody anticipates.
The first is control: who asked what and what was returned. The second is product insight, because a list of the questions people ask is the most reliable guide you will ever get to which reports are worth building and where your process problems are.
Practical Deployment Advice
Two things consistently determine whether this is adopted or abandoned after a fortnight.
Where ChatFin Fits
ChatFin answers from live JD Edwards data inside authenticated sessions that carry the requesting person's role, with the supporting rows attached to every figure, and delivers into the surfaces your people already use.
Because the substrate is the same regardless of surface, adding a second delivery channel later is a configuration rather than a project. Our page on conversational financial analysis covers the underlying requirements in more depth.
Choose the Surface for Convenience, the Platform for Substance
ChatFin delivers JD Edwards answers into Teams, a portal or a chat surface, with queries running inside authenticated sessions that inherit each person's role and every figure carrying the rows behind it.
A wrong number in a spreadsheet is one person's mistake. The same number pasted into a channel is organisational fact by tomorrow.
Frequently Asked Questions
Can employees ask JD Edwards questions inside Microsoft Teams?
Yes, and the surface is the straightforward part. What determines success is the layer underneath: whether the answer is produced by a real query against live data, whether the requester's own JD Edwards role governs what they can see, whether the supporting rows are attached, and whether everything is logged. That substrate is identical whether the answer appears in Teams, a portal or a chatbot.
How do permissions work when people have no JD Edwards login?
That has to be a deliberate decision. The clean design is that queries run inside an authenticated session carrying the requesting person's environment and role, with their chat identity resolved to their JD Edwards user so entitlement is inherited. For people with no JD Edwards account, decide explicitly what they may see, if anything, rather than defaulting to a shared identity with filtering bolted on.
What should we log?
Who asked what, the query executed and what was returned, which serves the control requirement. The second purpose is the one nobody anticipates: the list of questions people ask is the most reliable guide you will get to which reports are worth building, and repeated questions usually indicate a missing report or a broken process upstream.