Legacy to AI

We selected DB2 tables and the agent still cannot call them

We selected DB2 tables in Populate APIs and the agent still cannot call them. Did we miss a step?

One agent for a CoreIQ API and a Hub project, or one agent per product?

Can one Gene agent call a CoreIQ API and a Hub project in the same conversation?

Copy one table so the agent can practice off production DB2

We want the agent to practice on a copy of one table, not on production DB2. How do we move just that table?

Can Screener drive IBM i DSPF maps the way it drives BMS?

We assumed Screener could drive DSPF maps on IBM i the same way it drives BMS. Can it?

Which 'operation' does the agent actually run?

People keep saying the operation and meaning two different things. Which one does the agent use?

Can the agent subscribe to VSAM changes instead of polling?

The agent keeps polling a VSAM file for new records. Can it subscribe to changes instead?

TermIQ masks screen values. Does Gene do that too?

TermIQ masks screen values before a model call. Does Gene do the same thing, or do we configure that twice?

Turn a working SELECT into a tool without pasting SQL into the prompt

We have a SELECT that returns the policy header the agent needs. How do we make that a tool?

Two agent calls at once, and the second had nowhere to send the map

We fired two agent calls together and the second session had nowhere to send the map. Is the remote player single-session?

Live terminal webhook or end-of-day trail upload?

Should we point OL Terminal at a webhook, or keep uploading trail files at the end of the day?

Read-only IMS for an agent: Universal JDBC or native DL/I?

The agent only needs to read an IMS database. When do we use Universal JDBC, and when do we use native DL/I?

Is dynamic tool loading the same as deleting tools from the agent?

If we turn on dynamic tool loading, are we removing tools the agent is allowed to call?

The DataIQ query assistant is empty and will not explain a SQL error

The AI query assistant opens and says nothing useful, and it will not explain a failed statement. What is missing?

We have both CICS/BMS and IMS/MFS. Which remote path first?

We have CICS/BMS and IMS/MFS. Which remote path should we stand up first so an agent can call a map?

Flow creation stopped on unidentified screens

Hub flow creation stopped because screens were unidentified. We have hundreds of snapshots. What is the minimum we have to name?

A dataset read showed a service account, not the person who asked

An agent read a VSAM record and the audit showed a service account, not the person in the chat. Is that expected?

The token bill jumped after we connected a Hub project

After we connected a Hub project, the agent started sending every operation schema on every turn. What do we change?

Should the agent query DB2 in place or work from a copy?

The agent needs three DB2 tables. Should we copy them to Postgres, or query them where they sit?

The agent needs named fields and an AID key, not a terminal emulator

Our agent has to enter a customer number and press an AID key. We do not want it driving a 3270 emulator. Is OL Screener the right fit?

Turn a week of 3270 trails into a repeatable agent workflow

We have a week of 3270 trails. How do we turn one business journey into something an agent can repeat?

Populate from CICS found dozens of programs. Which ones become tools?

Populate from CICS listed far more programs than this agent should call. Do we publish all of them?

Call one CICS program from an agent without a 3270 session

We want an agent to start a single CICS inquiry. Do we point the model at the green screen, or at an API?