Leveraging AI inside MFT for customer value
MFT environments are growing in scale and complexity. So, we’re exploring where AI could deliver practical improvements that offer real value to our customers - without compromising their governance or security.
From RPA‑driven automation of partner lifecycle and compliance workflows, to machine learning that detects anomalies and forecasts capacity risk, and Gen AI that enables natural‑language insights into what’s really happening, the opportunity is significant.
We’re also examining controlled uses of Agentic AI for approved, human‑gated actions, and smarter change management that allows staged rollouts with automated rollback. While these ideas are exploratory, they’re grounded in real user needs and operational reality.
MFT environments are growing in scale and complexity.
So, we’re exploring where AI could deliver practical improvements that offer real value to our customers – without compromising their governance or security.
We’ve grouped the ideas by AI category, including RPA (robotic process automation), machine learning, Gen AI/LLMs, and Agentic AI.
- In the RPA category, we’re considering how AI could automate partner lifecycle workflows (onboard/offboard, rotate credentials, enforce retention), or support compliance by capturing evidence and reporting on workflows.
- In the machine learning category, we’re throwing around ideas like using AI to detect anomalies on transfer patterns, performance, and authorised behaviour to reduce false positives. Or for capacity forecasting (“will this environment handle X growth?”) and delivering early warning that the current capacity is being stretched.
- In the Gen AI category, we could utilise AI’s natural-language analytics capabilities so you can ask: “What changed?” “Why are failures up?”, or “Top causes of problems this week?”
- Agentic AI, of course, would enable the system to take action beyond rigid playbooks, but only with explicit controls in place. For example, handling supplier onboarding end-to-end workflow using approved steps with human approval gates where needed.
One area we’re definitely exploring is the potential to use AI to store up and roll out MFT configuration changes en masse at a pre-determined time. Why? Currently, whenever you make a configuration change in MFT, it goes live straight away. The moment you press ''save,’ the new setting is applied to the solution.
But what if you could store those changes (in effect, creating a crib sheet of changes) and send them all live at midnight? Then, if one of those changes doesn’t go according to plan, there’s an automated rollback to revert to the earlier setting: the change management element becomes automated – without you having to be present.
While all, some, or even none of these AI ideas may come to fruition, what’s important is that they’re forward-looking and of potential value to users. However, to move to the next stage, it’s essential that we ensure there’s a legitimate use case for these AI additions and that we’re sensitive to our customers' governance and security needs.
So, watch this space 👀.
Copyright © 2026, Axway, All rights reserved.