It's 3 a.m. You're staring at a critical SUM phase that should have finished hours ago. Your downtime window is slipping. The PMO is asking for updates. And somewhere in the thousands of logs flooding your screen is the answer you need — if only you could find it in time.
If you’ve ever managed a large-scale S/4HANA upgrade, this scenario feels painfully familiar. Traditional approaches, including relying solely on Basis expertise, transaction codes, and reactive troubleshooting, struggle to keep pace with the complexity of migrations today.
The game has changed. SAP upgrades have evolved from a purely technical challenge into an operational intelligence problem.
In this blog post, we cover:
Why SUM becomes an operational problem during an S/4HANA upgrade
What a real S/4HANA upgrade looks like under pressure
How AI support S/4HANA upgrades
What teams gain from AI-driven upgrades
Modern SAP S/4HANA upgrades introduce challenges that go far beyond traditional Basis administration, especially when execution depends on the Software Update Manager (SUM).
Overwhelming log volume across thousands of SUM phases: Preprocessing and shadow system phases alone generate thousands of logs. During a time-critical S/4HANA upgrade, no individual can realistically monitor, correlate, and analyze this data in real time.
Runtime estimates that rarely match production reality: Every SAP landscape behaves differently. Mock upgrade timings often break down in production when edge cases, custom code, or unexpected errors surface mid-execution.
Limited control under RISE with SAP: In RISE-managed environments, backend access is restricted during key SUM phases. Traditional troubleshooting methods no longer apply, forcing teams to rethink how they monitor and respond during an upgrade.
Behavior changes between SAP releases: SUM behavior can change subtly between S/4HANA releases, often without clear documentation. What worked during a 2023 upgrade may fail in 2025.
This is why reactive, expert-driven troubleshooting no longer works. Successful S/4HANA upgrades now require operational intelligence for continuous analysis, pattern recognition across past upgrades, and real-time recommendations that keep SUM execution moving forward.
Let’s look at two real-world scenarios from a recent 10-day S/4HANA upgrade that show why institutional knowledge alone is no longer enough, and why AI-powered analysis is becoming essential during high-pressure upgrades.
The ACT_UPG (Activation Upgrade) phase has long been one of the riskiest stages in an S/4HANA upgrade. This is where SAP activates the data dictionary before data migration. If objects fail to activate, the upgrade stops immediately.
Historically, Basis teams relied on a critical workaround. When activation errors occurred in unused components, such as legacy third-party add-ons scheduled for removal after the upgrade, you could accept the errors and allow SUM to continue. Cleanup could happen later. That assumption no longer holds.
With S/4HANA 2025, SAP removed the ability to accept certain activation errors during ACT_UPG. This change was not clearly highlighted in upgrade guides and was easy to miss, even for experienced teams. When these errors occur now, there is no workaround. SAP support’s only recommendation is to roll back, fix the issues on the source system, and restart SUM from the beginning.
If this happens six hours into ACT_UPG, your entire upgrade schedule collapses. A single undocumented behavior change can add days of downtime.
Lesson learned: Any activation issues found during a mock upgrade must be fully resolved in the source system before production. Past SUM workarounds cannot be assumed to exist in future S/4HANA upgrades.
Another scenario that frequently catches teams off guard during an S/4HANA upgrade is the loss of backup and restore access in RISE with SAP environments.
In traditional upgrades, teams routinely back up critical tables before attempting risky changes. If a fix fails, the backup is restored, and troubleshooting continues. It’s standard upgrade practice.
In RISE environments, that safety net may not exist. Customers do not have direct access to the HANA backend. Even if tables are exported or backed up using SQL tools, the files are written to the HANA server, an environment the customer cannot access. The result is a backup that cannot be retrieved or restored when needed.
The fix: This constraint must be addressed during upgrade planning, not during execution. Teams need to validate backup and restore capabilities upfront, coordinate with SAP RISE support, and design troubleshooting workflows that align with what is actually possible in a managed environment.
These scenarios reveal a fundamental challenge with modern SAP S/4HANA upgrades: the knowledge required to execute them successfully is fragmented. It lives across past projects, release notes, obscure SAP documentation, ticket histories, and hard-won individual experience. No single expert can retain all of it — and no team can search through it fast enough when an upgrade is under time pressure.
This is where combining tools like IT-Conductor’s SUMMon™ with agentic AI changes the upgrade experience. Instead of reactive firefighting, teams gain continuous operational intelligence throughout the upgrade.
All SUM phases are monitored as they run, with logs analyzed in real time. When anomalies occur, alerts include context — what failed, why it matters, and how similar issues behaved in past S/4HANA upgrades.
No more discovering at 3 a.m. that a phase expected to run for 30 minutes has been stalled for three hours.
Agentic AI continuously compares current SUM behavior against institutional memory from previous upgrades. Did the mock upgrade encounter this issue? How was it resolved? Is the current runtime deviating from historical norms?
This intelligence is surfaced automatically — without relying on tribal knowledge or manual searches through old documentation.
When errors occur, the AI goes beyond flagging failures. It analyzes the specific failure pattern and provides recommended next actions, often enabling teams to resolve issues in minutes rather than waiting hours or days for SAP support.
In practice, SAP support tickets during non-production upgrades can take hours or even days to progress. Teams upload dozens of logs, wait for analysis, and repeat the cycle. With AI-powered analysis, teams can often identify the likely root cause and resolution path within minutes, with a high degree of confidence, before external support even responds.
When S/4HANA upgrades move from firefighting to intelligence-driven execution, the biggest gains show up at the team level.
SAP took a step forward in the S/4HANA 2025 release by allowing multiple users to work in the SUM GUI at the same time. Previously, only one person could actively interact with SUM, while everyone else was pushed into a limited observer mode.
AI-driven upgrade intelligence goes further. Instead of gating visibility behind system access, the entire upgrade team can see real-time phase status, logs, error analysis, and recommendations from a single interface. Basis, application, infrastructure, and project stakeholders all work from the same source of truth without needing direct access to the SAP system.
This is especially valuable for:
Global follow-the-sun teams handing off upgrades across time zones
Cutover windows, where minimizing system access reduces risk
Faster decision-making when issues span multiple domains
Collaboration shifts from “waiting for updates” to shared situational awareness.
Every SUM phase, error, decision, and resolution is captured automatically as part of the upgrade process. There’s no need to reconstruct events after the fact or rely on incomplete notes once the pressure is over.
As soon as the S/4HANA upgrade completes, teams can generate comprehensive reports:
Which phases required manual intervention
What errors occurred and when
How each issue was resolved
Who took action, and at what point in the timeline
This documentation becomes a living runbook, not a static plan written months in advance, but an evolving knowledge base that improves with every upgrade cycle.
Instead of starting from scratch for the next migration, teams start smarter.
Upgrading to SAP S/4HANA like a boss means staying in control without scrambling when unexpected errors or undocumented SUM behavior changes surface. Your team isn’t troubleshooting in silos while the PMO waits for answers. Instead, you’re backed by institutional memory, real-time analysis, and intelligent recommendations that remove guesswork from the process.
AI won’t fully automate SAP upgrades, but it can dramatically shorten resolution time, surface risks earlier, and help teams stay ahead of problems instead of reacting to them. When downtime windows are tight and the stakes are high, that difference matters.
Watch our episode where I break down how AI is reshaping S/4HANA upgrades and share a preview of what we’re building with AI.