For most of my career working with SAP, one principle went unquestioned: change must be slow and careful. I actually lived that theory. After all, SAP runs the business: finance, supply chain, manufacturing, and revenue. When a change goes wrong, it could be a P1 – the kind of incident that stops operations and, sometimes, ends careers. 

To navigate that risk, the industry built a world around caution with heavy governance, long testing cycles, sequential delivery, and approval on top of approval. Out of it came one deeply held belief: that speed and safety are opposites. Go faster and accept more risk, or reduce risk and slow everything down – but gain confidence that nothing would break. That trade-off naturally became the default operating model for SAP change. 

For a long time, it was the right model. I now think it has become the most expensive assumption in enterprise IT, because the world it was built for is gone. 

The question I’d encourage any SAP leader to ask today is, “What is it costing us to keep changing this slowly?” 

A new pace of business change 

Innovation, driven by AI, is now moving faster than enterprise systems were ever designed to handle. This isn’t ABAP pace, or even traditional enterprise-software pace, but rather AI pace. Product teams iterate in days, and AI is compressing idea-to-execution from months into weeks. Yet I’ve seen large enterprises running legacy change processes report SAP change lead times of three to nine months. 

We call that distance the Agility Gap. It’s the growing gap between how fast the business needs to move, and how fast change can actually be delivered. My colleague Matthew Stammers has made the economic case for closing it and presented the four new metrics SAP leaders need to lead that conversation. 

I want to make the technologist’s case: why agentic AI can finally close the Agility Gap. 

The trade-off was never a law of nature

Across the wider software industry, speed and safety stopped being opposites years ago. DevOps proved it: the best teams deploy more often, not less, shipping small changes continuously rather than batching them into large, risky releases, resulting in fewer failures and faster recovery. The DORA research has shown it year after year. The root causes of slowness and fragility turn out to be the same things: manual steps, limited visibility, bundled releases. So, fixing them improves both at once. 

The best teams don’t move quickly by trading away safety. They move quickly because they’re confident in every change they make. 

SAP has been the exception, for understandable reasons: these environments are complex, tightly integrated, and carry years (or even decades) of customization that makes even a small change hard to isolate. Visibility is often limited, testing is manual, and consequences can be immediate. Faced with that, the industry defaulted to caution, and over time caution hardened into friction.  

The problem was never a lack of appetite to move faster; it was a lack of confidence in what each change would do. That isn’t a flaw in SAP, but a limitation of the old methods we wrapped around it – and precisely what agentic AI removes. 

From automating execution to an intelligent lifecycle 

For two decades, automation in SAP change meant taking a process we already understood and running it faster: scripting the transport, automating the regression pack, routing approvals. It helped, but it stayed confined to executing the steps we’d already mapped out. The judgment-heavy work of deciding which change to make, what it would affect, what to test, and why something broke, still sat with people and that was the real bottleneck. 

Agentic AI ends that constraint. An agent takes a goal, works out the steps itself, uses the tools around it, and acts, adapting as it goes. It’s the difference between a macro and an intelligent assistant. It affects every stage of the journey a change makes, from a business ask to live in production: design, impact assessment, build, test, release, and run. 

The right change starts with design

It starts at the front, with the stage we used to treat as paperwork: change design, where you decide what to build and whether it’s worth building at all. This is the highest-leverage stage, because a change designed well from the outset gives teams confidence in everything downstream, so it moves faster, with far fewer issues. The rework and late defects we blame on delivery often originate in a design that was ambiguous from the start. 

The old goal was to do change right. The new goal is to do the right change, in the right way, for your specific circumstances, at the pace your business needs. 

At the design stage an agent can create requirements, explore the real options for meeting a business need (standard configuration versus custom build, process change versus system change), test feasibility against the system as it actually is, and attach a credible estimate of cost, effort and return to each. Work that used to take weeks of consultant time (or never happened at all) now happens in the room while the decision is still being made. 

What changes downstream

The same intelligence runs through the rest of the SAP change lifecycle. Impact assessment that once took a senior consultant days of tracing dependencies across a customized estate, and still missed things, an agent can now do in minutes, reading the actual system rather than relying on what the documentation claims.

Orchestrating delivery across a complex landscape like sequencing changes, catching conflicts, and handling retrofits becomes something the system manages, rather than something a person has to manually track in a spreadsheet. In production, issues are caught and rolled back in minutes, instead of being discovered by a user and escalated into a P1. 

Testing also changes completely. Agentic AI won’t simply make today’s testing faster; it will make it a different activity altogether. The fixed regression pack, run out of habit and largely by hand, gives way to agents that work out what to test from the change itself and the system’s real behavior, generate and run those tests, and investigate each failure to its cause. 

The impact on the SAP change function 

The impact shows in the numbers that have defined the function for years, as well as one more that was never on the original scorecard: 

Volume. The change function has always been capacity-bound: only so many senior people, only so many changes they can properly analyze, test and shepherd to production. Agentic AI lifts that ceiling, so the same team absorbs far more change — S/4HANA programs, AI-driven process redesign, regulatory deadlines, business-as-usual — without the backlog throttling the business. 

Velocity. When teams know what a change will affect before it’s built, the slow, judgment-heavy steps that widen the Agility Gap – impact analysis in days, test cycles in weeks – stop being points of hesitation. Agents compress them to minutes and hours, Time to Value moves from months toward weeks, and the business stops waiting three quarters for a new capability. 

Defects. Since impact is understood before a line is written, and testing is comprehensive and continuous rather than sampled and manual, more change no longer means more failure. It actually means less failure. Failure rates fall, recovery is measured in minutes, and the fastest teams become the safest – exactly as DORA predicts. 

Alignment. This is the gain that was never originally on the scorecard: because options, cost, effort and return are surfaced at design in terms both sides understand, agentic AI gives business and IT a shared language. IT stops receiving requirements it can’t question; the business stops receiving estimates it can’t interrogate. Together they prioritize the portfolio by value rather than by whoever argued hardest — and for a function whose hardest problem has always been deciding what to do, not how to do it, that may be the most valuable shift of all. 

The role of the SAP professional changes with it: from doing the manual analysis and testing, to directing and supervising the agents that do. They move from operator to orchestrator, with human judgment moving up the chain to focus on which changes are worth making and when to overrule the machine. 

The evolution from caution to confidence 

This is the thinking behind our approach at Basis Technologies: Intelligent Change Management, an operating model for managing SAP change with the new generation of capabilities, including agentic AI. It helps teams design changes more accurately, understand their impact earlier, and orchestrate delivery across complex landscapes with confidence, moving from a business request to live in production faster, with fewer defects and full control.

That’s what it means to change intelligently. Our software sits alongside the platforms SAP customers have already invested in to extend what they do, rather than replacing them. 

The takeaway for enterprise change leaders 

Delivering change safely will always matter — it’s the floor — but designing and delivering the right change at the pace the business now moves is the ceiling, and the ceiling is where competitive advantage lives. 

The rules of change have been rewritten: with confidence in every change, speed and safety are no longer a trade-off. The work now isn’t just to execute change well, but to choose and shape the right change in the first place. Agentic AI is what now makes this possible at scale. 

SAP can support change at pace; that is settled. The question is whether your change model gives your teams the confidence to match the pace of your business — and the organizations that close that gap will define the next phase of enterprise performance.