Help us shape Aida Conversation Mode - Early testers wanted
We’re opening up a new capability to a small group of early testers, and we want to be upfront: it is not finished yet.
That is exactly why we need you.
What is Aida Conversation Mode?
Many of you already use Aida or know her through AI Assist, the scripting assistant that helps generate and refine Exalate scripts.
Conversation Mode starts one step earlier.
Instead of beginning directly with scripts, you begin with your integration requirements. You describe what you want the integration to do. Aida asks questions, captures and structures your requirements, and translates them into the underlying Exalate scripts as the conversation progresses.
The goal is to make powerful integrations easier to define, test, and adapt, starting from intent rather than code.
The direction behind Conversation Mode
This is an early step in a broader direction we’re building toward: making integrations easier to describe, shape, and refine, while giving people with workflow and functional expertise a more direct role in defining sync logic — without sacrificing the precision, control, and governance that complex cross-company syncs require.
What the early version already supports
-
Starting from requirements, not scripts: describe the expected behavior in plain language
-
Continuous script generation as the conversation progresses
-
A Build/Test loop: switch between configuring and validating at any point
-
Iterative refinement: pick up where you left off, correct behavior, and refine further
-
Inline feedback directly on any Aida response in the conversation
What doesn’t work yet — know before you test
This is a feature-in-progress, not a finished capability.
We’re being transparent about the current gaps:
-
Trigger logic is not captured through conversation yet. Only manual syncing is supported in this release.
-
Script view may briefly show outdated scripts while Aida is still processing.
-
Local connections have not been fully tested. Please proceed carefully if that is your setup.
-
Aida does not always ask enough follow-up questions, which can lead to ambiguous requirements.
-
Script generation is still improving. Hallucination of non-existing Exalate API methods has been observed.
What is coming next
We are actively working on:
-
trigger support
-
manual script corrections that do not break Conversation Mode
-
ticket-system metadata awareness, including custom fields, statuses, and transitions
-
automatic test-case generation
-
improvements to the quality of Aida’s follow-up questions and generated scripts
What kind of feedback we need
We’re especially interested in feedback from partners and customers who can test this against real scenarios.
Useful feedback includes:
-
Where did Aida understand the requirement well?
-
Where did it ask the wrong question or not enough questions?
-
Where did the generated script match the expected behavior?
-
Where did the script fail or become unclear?
-
Was the Build/Test loop useful?
-
What would make this easier to use in a real project?
How to get access
Conversation Mode is behind a feature flag in Exalate 4.0 and enabled per organization.
If you’d like to join the early tester group, reply here or reach out directly at product@exalate.com. We’ll share the guide and, if useful, set up a short walkthrough call.
This is not a launch announcement.
It is an invitation to test, break, question, and help shape what comes next.
Thank you!