Christian Klein just told Fortune magezine at Davos interview that keyboards are nearly done in SAP.
If you’re hiring SAP talent right now, this changes everything.
Klein isn’t talking about some distant future.
He’s talking 3–5 years.
Voice commands. Natural language queries. AI agents that traverse financials, supply chain, and sales to surface insights and trigger workflows.
𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻𝘀 𝗳𝗼𝗿 𝘆𝗼𝘂𝗿 𝗦𝗔𝗣 𝘁𝗲𝗮𝗺 𝗮𝗻𝗱 𝗵𝗶𝗿𝗶𝗻𝗴 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆:
Transaction codes become less valuable than the ability to frame business questions.
“Show me plants where inventory is misaligned with demand and recommend fixes” replaces `/nVA01`.
The next generation of SAP power users won’t say “I can build any ALV report.”
They’ll say “I can orchestrate AI agents across finance, supply chain, and sales to change how this business runs.”
Horizontal AI wins.
The companies that succeed won’t just add copilots to one module. They’ll let agents work across PowerPoints, spreadsheets, and SAP tables to connect the dots.
Geopolitics enters your architecture decisions.
Data sovereignty, geo-lock, local clouds, and portability move from CIO slide decks into daily conversations.
𝗜𝗳 𝘆𝗼𝘂’𝗿𝗲 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗼𝗿 𝘀𝘁𝗮𝗳𝗳𝗶𝗻𝗴 𝗦𝗔𝗣 𝘁𝗲𝗮𝗺𝘀 𝘁𝗼𝗱𝗮𝘆, 𝗵𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝗻𝗼𝘄:
• Hire for end-to-end process knowledge (order-to-cash, plan-to-produce) over single-module expertise
• Look for people who can design prompts like they’d design Fiori apps: clear intent, business context, measurable outcomes
• Add “data residency and cloud portability” to your must-have vocabulary for architects
The uncomfortable reality?
If your team still defines their value by “knowing where to click,” an AI agent will eventually out-click them.
How ready is your SAP landscape – and your hiring strategy – for a world where people talk to the system instead of type into it?
Related reading
Meet Joule: SAP’s bet on the conversational enterprise
SAP Joule is the company’s AI copilot, and it is the clearest signal of where this is going. Joule sits across the SAP suite: S/4HANA, SuccessFactors, Ariba, and the Business Technology Platform. Ask it a question in plain language, and it navigates, retrieves, and increasingly acts: creating purchase requisitions, summarizing a supplier’s risk profile, drafting job descriptions from data scattered across modules.
The detail that matters for hiring managers is the rollout order. Joule arrived first in SAP’s public cloud offerings, then expanded to private cloud. That means the customers moving fastest toward the conversational model are the ones on RISE and GROW, on clean core, on current releases. Your AI readiness is now directly tied to your cloud and upgrade posture. A heavily customized ECC system cannot talk to Joule, because Joule has nothing clean to talk to.
From copilots to agents: the real shift
A copilot answers your questions. An agent does your work. That distinction is the whole story of the next five years.
The agents coming to SAP will traverse financials, supply chain, and sales the way a good analyst does: noticing that inventory is misaligned with demand in three plants, tracing it to a forecast error and a supplier delay, drafting the corrective purchase orders, and presenting you with a decision instead of a spreadsheet. Procurement agents will handle routine buys. Finance agents will draft the month-end commentary before the controller opens the file.
This is why horizontal AI wins. A copilot trapped inside one module is a nicer search box. Agents that work across SAP tables, spreadsheets, email, and documents are a different operating model. The companies building that muscle now, with clean data and clear process ownership, will compound the advantage.
What changes for functional consultants
The functional consultant’s job is not disappearing. It is moving up the stack.
Knowing the IMG tree by heart becomes less valuable than knowing the process cold and describing a desired outcome precisely enough for an agent to configure it. The consultants who thrive will be the ones who can validate AI-generated configuration, spot the hallucination in a pricing procedure, and explain why the agent’s proposed credit management setup will break in your Latin American entities.
New specialties are already emerging. AI trainers for SAP: functional experts who teach agents the company’s processes, terminology, and exceptions. Agent supervisors: people who review what the agents did overnight the way a manager reviews a junior analyst’s work. If you are hiring functional consultants today, hire for judgment and process depth, not transaction speed.
What changes for end users
For twenty years, SAP power users defined themselves by transaction literacy: knowing the T-codes, the workarounds, the exact click sequence that makes the system do what finance needs. That entire skill set is being automated.
What replaces it is process literacy and exception handling. The user of 2030 does not memorize transactions; they describe outcomes, review what the agent proposes, and intervene where judgment is required. The super-user evolves into something more valuable: the person who knows the process deeply enough to catch the agent’s mistakes and teach it the exceptions that never made it into the documentation.
The blunt implication for training budgets: stop training users on clicks. Start training them on the business process, on data quality, and on how to interrogate an AI’s output. A user who trusts the agent blindly is a controls failure waiting to happen.
The risks nobody is pricing yet
Every agent that posts a journal entry, releases a purchase order, or changes a delivery schedule is acting inside your system of internal controls. Segregation of duties was designed for humans. Audit trails were designed for humans. When an agent executes a hundred transactions overnight, who approved them? Who is accountable when one is wrong?
Then there is data residency. Agents that move your data across borders create exposure under regimes like the EU AI Act and a growing set of data sovereignty rules. Where the agent runs and where the data goes is now an architecture question, not a legal footnote.
The uncomfortable reality from the original post stands, with one refinement. If your team’s value is knowing where to click, an agent will out-click them. If your team’s value is knowing how the business runs, no agent can replace that, and every agent you deploy will make them more powerful. Hire accordingly.
