Muhammad Alam opened TechEd 2025, by saying that developers are not being replaced by AI, but they are being upgraded by it. He framed the moment as the start of an “agentic AI” era, where software doesn’t just suggest next steps but pursues them with guardrails.
The message was that developers can still expect to write code, but also design workflows and supervise agents that deliver outcomes across finance, supply chain, HR and CX. There were many announcements, from data fabric updates to new runtime and tooling choices on SAP BTP.
Alam described the evolution of the portfolio, with SAP BTP as the engine, and the Business Transformation Management portfolio as the navigator, connecting strategy, process and people. Technical upgrades are now directly tied to delivery work: process redesign on Signavio and WalkMe now sit next to agent design on BTP, not in a separate track.
The practical changes fell into two groups.
First, SAP Build now integrates more smoothly with the other SAP tools already used. Local Model Context Protocol (MCP) servers let popular coding agents and IDE companions, such as Cursor, Windsurf, Claude Code and Cline, talk to SAP frameworks. A new SAP Build extension for Visual Studio Code allows Build capabilities to be accessed in a familiar environment. SAP also trailed plans with n8n so Joule Studio agents can cooperate with n8n agents, which is useful if you’re wiring SAP events to non-SAP automations. For mixed teams, this lowers the friction of moving between low-code assembly and pro-code extension work.
Second, Build itself has become an agent workshop. Joule Studio adds agent-building tools that let you create task-specific agents grounded in SAP data and context. These can act based on changing business conditions rather than waiting for a human to click a button. If you’ve been asked to “add AI” to an extension, this is the path that keeps governance and data context intact across the stack.
Every AI demo lives or dies on data quality and SAP illustrated that point by extending Business Data Cloud in two directions. One is connectivity, with a new solution extension with Snowflake, plus Business Data Cloud Connect for Snowflake, which joins earlier integrations with Databricks and Google Cloud. For architects, the appeal is zero-copy sharing and the chance to keep workloads on the platform that fits cost and performance, without breaking governance or losing business context. This is the kind of integration that lets analytics and agents link SAP and non-SAP data without expensive duplication.
The second is modeling. A new data product studio in Business Data Cloud turns curated sources into “data products” ready for analytics, AI and apps. HANA Cloud’s knowledge graph also grows, with automatic graph generation across tables and models so builders can see relationships rather than isolated tables.
The keynote sketched an AI stack with three layers: models, orchestration and operability.
At the model layer, SAP introduced SAP RPT-1, a relational foundation model built for structured business data. It predicts business outcomes such as delivery delays and payment risk, and SAP opened a free playground for developers to experiment. This sits alongside Joule’s growth from helper to coordinator, with new assistants that can plan and complete tasks across departments by directing multiple agents.
At the orchestration layer, SAP talked about an agent-to-agent protocol to let agents from different vendors work together, which matters if your landscape mixes SAP agents, hyperscaler services and third-party tools. It also points to a future where “agent handoffs” are normal in long-running business processes.
SAP BTP is being pitched as the place to run these agents natively (in line with their Clean Core initiative), rather than as sidecars bolted onto apps. If you followed the “AI Foundation” thread from Sapphire, the picture now includes an AI Agent Runtime and a prompt optimizer to harden solutions and reduce prompt adjustments.
A repeated theme was providing for developers in their preferred environments.
The VS Code extension for Build and local MCP servers make a noticeable difference on day one because they let you keep established workflows while tapping SAP frameworks.
ABAP teams get a bridge into Build, with visual service creation and UI wiring that pairs clean-core goals with practical delivery speed. SAP also flagged a future ABAP-based MCP server to expose ABAP Cloud development and ABAP AI features to coding agents, which is relevant if you’re standardizing on agentic assistants in the IDE.
There was also a people story. SAP committed to help 12 million people gain AI-ready skills by 2030 through expanded training and certification, including work with Coursera.
For consulting firms, that is a signal to build learning plans that mix BTP services, Build, data modeling on Business Data Cloud and agent design. In practice, the most valuable consultants over the next year will be the ones who can translate process pain into an agent plan, wire the data products, and deliver a small, reliable automation.
Taken together, these changes make for a clearer development path. If you build extensions, you now have native ways to create and operate agents that act across SAP and non-SAP systems, with Build and BTP providing the assembly and runtime layers. If you focus on data, Business Data Cloud’s connectors and modeling tools make cross-platform scenarios more practical without copying data everywhere.
If you run delivery teams, the skill pledge and the move toward standard protocols suggest that agent work will spread across projects. The keynote avoided hype and drilled into what you can ship this quarter: better tooling, a stronger data layer, and AI that can undertake supervised action.
Aside from the keynote, SAP released a broad set of updates that point to how AI and automation will shape daily work across its ecosystem. The details are dense, but the key message is clear: nearly every SAP domain is gaining some form of agentic intelligence, data connectivity, or guided automation.
In finance and procurement, SAP is moving from manual processing to AI-supported decision-making. Task-based agents handle reconciliation and classification work, while the next generation of Ariba brings AI into contract reviews and sourcing. For consultants, this means treating finance and spend workflows as adaptive systems that evolve with embedded automation rather than static processes.
In workforce management, Fieldglass is shifting from title-based searches to skills-based matching, reflecting how project staffing now depends more on capabilities than job roles. For delivery teams, it’s a signal to organize around skill visibility and continuous resourcing rather than fixed assignments.
Across the supply chain and business networks, SAP is strengthening integration between ERP, supplier collaboration, and logistics. New features in SAP Business Network and Cloud ERP reduce friction in supplier discovery, disputes, and sustainability reporting — all supported by Joule as an intelligent assistant. Developers should expect supply chains to become event-driven environments where data and prompts flow freely between systems.
The integration and operations layer is becoming more intelligent. SAP Integration Suite gains tools for AI-based monitoring and exposes APIs as callable services for agents through the Model Context Protocol. Integrations now function as living components of automation, not just plumbing.
In transformation and architecture, Signavio, LeanIX, and WalkMe are connecting more tightly, turning process, architecture, and adoption data into one loop of continuous improvement. Architects will spend less time documenting change and more time designing adaptive systems where processes and agents evolve together.
Finally, project delivery gains structure through updates to RISE with SAP and Cloud ALM. These add AI-guided project paths and preconfigured templates, reinforcing the principle: keep the core clean, and capture business-specific logic on BTP.
Taken together, these updates describe a single direction. SAP’s tools are converging around automation that acts with context: not just data processing but guided decision-making. For consultants and developers, the key point isn’t to memorize each feature, but to watch where agents, data, and design now overlap, and build projects that make use of those intersections.
So, the boundaries between data, process, and automation are blurring, and success will depend on understanding how these layers now interact.
Consultants should start treating SAP BTP as the natural home for adaptation and experimentation if they don’t already. The platform is where agents, integrations, and data products converge, and where custom work can live without disturbing the clean-core principle. Familiarity with its services including Build, Integration Suite and Business Data Cloud is increasingly important in maintaining relevance, though the timeline for these technologies to be widely adopted will take longer than SAP would like.
Developers, meanwhile, will need to grow comfortable working alongside intelligent systems. The arrival of Joule Studio, Model Context Protocol (MCP), and agent runtimes means code will increasingly be consumed by, and written for, AI-driven tools. Understanding how agents access data, invoke APIs, and observe governance constraints will soon be as important as traditional ABAP or CAP skills.
At a broader level, data fluency is becoming essential. The new Business Data Cloud and its modeling tools are designed for reusable data products, and developers who can model, expose, and govern data in this way will find themselves central to every AI-driven initiative.
For consulting practices, these shifts call for cross-functional awareness. Process specialists will need to grasp how agents interact with workflows, architects will have to account for data movement and protocol compatibility, and developers will design solutions that operate through orchestration rather than direct execution. The work becomes less about coding isolated components and more about designing how intelligent systems collaborate safely and effectively.
Finally, it’s worth investing in continuous learning. SAP’s expanding training ecosystem and new AI-focused certifications provide the structure, but the real preparation happens through experimentation — testing small, low-risk use cases that explore how agents, data models, and integrations behave in production settings.
In short, the consultants and developers who excel in this next phase of SAP services will be those who understand both how SAP’s technology is changing and how to evolve their own working habits: moving from system maintenance to intelligent system design.
If you are an SAP professional looking for a new role in the SAP ecosystem our team of dedicated recruitment consultants can match you with your ideal employer and negotiate a competitive compensation package for your extremely valuable skills, so join our exclusive community at IgniteSAP.
Python for SAP Consultants: Where It Appears and How Much You Need
