AI has spent the last few years bouncing between hype and experimentation. For many businesses, it felt promising, but not yet practical. That is starting to change.
Watch the webinar on this topic
Want the full walkthrough and practical examples? This webinar breaks down what’s real, what’s new, and how to move forward with AI in 2026.
Over the last year, AI moved from being interesting to being usable. Models got better at following instructions, responding faster, and reasoning through more complex tasks. Multimodal capabilities went mainstream. Grounding answers in company data became more practical. And, just as importantly, AI started showing up inside the enterprise software people already use every day.
That changes the conversation. The question is no longer whether AI matters. It is where it is already creating value, how vendors like SAP are embedding it into core business workflows, and what organizations should do next if they want outcomes instead of experiments.
A few shifts came together at once.
Models improved in meaningful ways. Better instruction following, faster responses, stronger reasoning, and more capable agent-style behavior made AI far more useful for real business work. Open-source models also closed part of the gap, expanding what companies can do and making experimentation more accessible.
Multimodal AI went mainstream. Businesses do not operate on clean text alone. They work with emails, screenshots, invoices, PDFs, reports, meetings, audio, and video. AI can now work with much more of that real-world, unstructured information than it could even a year or two ago.
It became easier to ground AI in enterprise data. Search, retrieval, and integration tooling matured quickly, making it more practical to get answers based on company context rather than generic knowledge.
Embedded AI, agents, and governance started arriving together. AI is now being built into core business suites, while agent capabilities are moving beyond question-and-answer into multi-step task execution. At the same time, governance is catching up, giving organizations better ways to manage security, oversight, and risk.
Key takeaway: Companies are moving from AI exploration to AI-enabled business outcomes.
Organizations do not have to imagine the use cases anymore. They can already see them in the enterprise software they use today.
In customer service, AI is helping teams understand cases faster, determine likely next steps, and surface relevant resolutions so issues can be resolved more consistently.
In sales, AI can pull together client context from multiple sources, highlight what matters, and help teams show up better prepared for conversations.
In marketing, AI is helping teams create, edit, adapt, and personalize content across channels, formats, and audiences without rebuilding everything from scratch.
In operations, AI is reducing the time spent searching for information, navigating disconnected steps, and piecing together fragmented workflows.
In finance and accounting, AI is supporting document extraction, issue detection, invoice analysis, and exception handling, which helps cut manual effort and reduce errors.
In reporting and analytics, users can increasingly ask questions in plain language, investigate what is driving the numbers, and get answers that previously required manual analysis.
The point is not that every use case is fully mature. It is that AI is already landing inside the suites companies use across departments, and that is where measurable impact starts.
SAP has been clear about its direction: AI, combined with business data, is reshaping how companies interact with systems and how work gets done.
That strategy shows up in five priorities.
Joule becomes the new AI user experience. SAP is evolving Joule from a chat-style assistant into a more capable digital work assistant that can answer questions, support tasks, recommend actions, and increasingly help orchestrate work.
AI gets embedded and extended across business processes. Rather than keeping AI separate from the system, SAP is placing it inside the workflows where users already work. Customers and partners will also have more ways to extend those capabilities with low-code, no-code, and pro-code tools.
Industry processes get reimagined. SAP is applying AI to specific industry workflows, not just generic productivity use cases, so processes can be redesigned around better decisions and more automation.
Business Data Cloud becomes the foundation. SAP is investing in curated data products and stronger harmonization across SAP and non-SAP data so analytics and AI can run on better business context.
Migration and development get AI support too. SAP is also bringing AI into the work of consultants and developers to help reduce the cost and effort of moving customers to cloud solutions.
For organizations on SAP Business ByDesign or SAP Business One, many of these capabilities do not apply directly to today’s product footprint. But they do offer a useful preview of the direction SAP Cloud ERP is heading.
For years, work in business software meant navigating dozens of applications, searching for data, entering transactions, and manually stitching processes together. SAP’s vision with Joule points to a different way of working.
Instead of jumping across systems to complete every step manually, users will increasingly interact through natural language. They will ask questions, surface insights, trigger actions, and coordinate work across systems more naturally. In that model, AI is not a separate destination. It becomes part of how the work gets done.
That does not mean AI replaces people. It means human workers and digital workers will increasingly operate together, with AI handling more of the repetitive, cross-system, and insight-heavy tasks while people stay focused on judgment, review, and decisions.
This shift will not happen overnight. But the direction has already been set, and SAP is already building toward it.
SAP’s advantage in AI is not just the model layer. It is the combination of applications, data, and AI working together.
Its applications run mission-critical processes across finance, supply chain, procurement, sales, operations, and more. Those processes generate semantically rich business data, meaning the data carries business context, structure, and process meaning. That is exactly the kind of foundation AI needs to be useful in enterprise settings.
This is where the flywheel becomes important. The applications generate valuable business data. That data improves AI. The AI gets embedded back into the applications. Over time, that creates compounding value.
In other words, the future of ERP is not just better screens. It is a combination of trusted applications, connected data, and embedded AI working together.
SAP is not just talking about AI strategy. It has already delivered a wide range of AI capabilities across its portfolio, including many inside SAP Cloud ERP, with more on the roadmap.
A few examples stand out.
Joule across cloud solutions. Joule is already available across SAP’s cloud landscape, and its role is expanding with each release.
Sales order support. AI can help extract sales order data from customer purchase orders and reduce the effort and errors that come with manual entry.
Fulfillment issue detection. When orders go off track, AI can help surface problems faster and guide users toward resolution.
Billing and AP issue analysis. AI can analyze posting issues on customer invoices and supplier invoices, helping teams fix blockers more quickly.
Natural-language reporting. Users can ask questions of reports in plain language, drill into what is driving changes, and get summaries that are easier to understand and communicate.
Emerging agents for finance. Newer capabilities point toward agents that can analyze approvals, draft postings, monitor signals, and propose scenarios so finance teams spend less time gathering information and more time deciding what to do.
A practical note on licensing: SAP has worked to simplify how AI capabilities are packaged. In general, base AI covers included capabilities already available in SAP cloud solutions, while premium AI applies to more advanced Joule, agent, or consumption-driven use cases. That makes it easier to start with what is already available and expand intentionally.
Most organizations follow a similar AI journey.
At the beginning, there is experimentation. Teams test tools, learn quickly, and start to see possibilities, but the work is often fragmented and the business value is hard to measure.
Next comes disciplined piloting. This is where organizations prioritize use cases tied to strategic goals, run focused pilots, and make faster decisions about what should scale and what should stop.
Then comes measurable value. AI becomes embedded in real workflows, adoption expands, and the business starts to see tangible outcomes such as time saved, lower costs, improved decisions, stronger consistency, or reduced risk.
Finally, organizations move into scale. At this stage, success is less about isolated wins and more about repeatability: shared data foundations, reusable integrations, stronger governance, and delivery approaches that support broader rollout.
No matter where a company starts, the path forward looks similar: move to the right by delivering measurable outcomes inside the workflows that matter most.
There is no shortage of excitement around AI. The challenge is turning that excitement into a repeatable operating model. A practical starting playbook looks like this.
Lead with the business problem. Start with the process pain, decision bottleneck, or business outcome that needs to improve. Let that define the solution, not the other way around.
Start with what already exists, and buy proven before building custom. Many enterprise platforms already include AI capabilities teams can use right now. It is often faster and lower risk to create value there first, then reserve custom development for the areas that are truly differentiating.
Start small and make the win measurable. Keep the scope focused. One workflow. One decision point. One outcome. Deliver the win, prove the value, and expand from there.
Build reusable foundations as you go. Each initiative should strengthen the data, integrations, governance, and delivery methods the next initiative can reuse.
Make change and adoption part of every deliverable. Even a strong AI solution will underperform if users do not understand it, trust it, or adopt it.
Organizations that do these things consistently are the ones most likely to move from isolated experiments to durable business value.
AI for business is no longer a future concept. It is already reshaping how people research, decide, create, analyze, and execute work across the enterprise.
What changed is not just the quality of the models. It is the combination of better models, multimodal capability, easier access to company-specific context, embedded AI inside enterprise applications, and stronger governance.
For SAP customers, that direction is becoming increasingly concrete. The strategy is clear, the roadmap is active, and the practical question is no longer whether AI belongs in the business. It is where you can apply it first to create measurable outcomes.
VistaVu helps teams identify the right AI use cases, strengthen data foundations, and apply SAP’s embedded AI capabilities to deliver measurable outcomes.
Prefer email? Reach out and we’ll connect you with the right VistaVu expert.