2026: The Year AI Gets Real, and Workflow Orchestration Takes Center Stage

Now that we have made it past the first month of 2026 after a year 2025 that was full of AI hype, experiments, some failures, and even some successes, it is time to take stock. More importantly, it’s time to look forward to what we can, and should, do better in this and the years to come.

2025: AI Exploration and Experimentation

One thing that became abundantly clear in the course of 2025 is that LLMs and their offspring, digital agents, are not (yet?) the silver bullet that vendors and some analysts wanted us to see them as. On the other hand, we have learned that there is considerable potential in these technologies if they are implemented and applied wisely.

Which, in general, so far, we have not. Although businesses started to wake up. Several research projects, including TheAgentCompany, CRM Arena Pro, or The State of AI in Business 2025, informed us that complicated, not even talking about complex processes, cannot yet be run autonomously. According to Wharton, we have left the stages of exploration and experimentation behind and can see value and ROI in generative AI implementations. And, we might now see acceleration and perhaps an inflection point.

What seems proven at this time is that digital assistants work and do deliver value, while (autonomous) agents don’t do this so much, although we see limited successful scenarios, in particular in the customer service area.

There is hope that the current hype cools down and gets replaced by some careful realism, not only on the buyer side, but also, or rather especially, on the vendor side. Bluntly, the grand marketing stories about the agentic enterprise now need to be backed up by some serious delivery on the workflow and orchestration level. This is especially true after Salesforce’s semi-successful move of its support organization to Agentforce,

At this time, it appears that workflows and cross-application orchestration are the next, or perhaps even the current, battleground of enterprise software vendors. The focus shows. Every major vendor is adding agentic capabilities to their cross-application layers, whether they are as big as Microsoft, Oracle, Salesforce, SAP, or ServiceNow, or smaller companies like Zoho, Creatio, or Pega. Not to speak of companies with no-code, workflow, and automation roots like UiPath and its competition.

Combine this with the ongoing struggle of dominating the gateway to the enterprise user, and we are looking at some interesting years to come.

These will be some years in which business leaders will have to take some decisions that need very careful deliberation in order to not get trapped on either of these frontiers.

The 2026 Battleground

This is because not the smartest model gives the decisive edge. Let’s get real here, AI models are a commodity. Neither enterprise leaders nor users will really care about which one is in use at any given time, but they will expect them to work seamlessly and flawlessly in the background.

The real edge for vendors is where users’ needs and habits meet context and where there is a short path from intent and insight to action. This is on the frontend- and the workflow level.

What is clear is that going forward, access to applications via the application frame is not the only way, probably not even the most important one. With MS Teams and Slack, supported by their respective agent or copilot ecosystems, Microsoft and Salesforce are very well positioned and fighting to dominate this front door to the applications. The main competition here is perhaps Google, although Google does not offer business applications. Still, many businesses, especially SMBs, are running Google Workspace, which puts Google Chat and Gemini on the scene.

Vendors that own this space will have a strong position in delivering and governing digital assistants, akin to the bouncer at the door of a club. This position also keeps open the door to cross-application workflow automation, which will be the other big fight of 2026 and potentially beyond.

And there is far more competition. In the Titan league of enterprise software vendors, we find the likes of Microsoft, Oracle, Salesforce, SAP, and ServiceNow. All of them are building agentic and orchestration layers. All of them claim support for multi-agent scenarios and cross-vendor support. All of them come from a different starting position that reflects their respective area of strength.

• Microsoft bets on the ubiquity of M365 and Dynamics on the base of Azure.
• Oracle doubles down on its infrastructure play via OCI, combining hardware and business apps with the strength of its database engine.
• Salesforce leverages its dominant CRM position, Slack, its integration capabilities via MuleSoft, and its strong share of mind.
• SAP works from the moat of its strong ERP position with its exhaustive industry process knowledge and a heck of a lot of business transactions touching an SAP system.
• For ServiceNow, everything starts from the workflow itself, which can position the company as a “neutral entity” although it, too, offers business applications.

And then, outside of the titans, there is the whole gamut of smaller, yet viable, enterprise software vendors. There are smaller suite vendors like Zoho, or others that come from either the workflow corner, like Pega or Creatio, the RPA arena like UiPath, or have a low-code and no-code legacy. Plus, there are the customer service and contact center specialists who built and/or acquired corresponding capabilities, too.

In summary, most significant vendors offer orchestration capabilities.

What Enterprise Technology Decision Makers Should Consider

Therefore, there is no point anymore in asking which vendor offers AI orchestration. It has simply become a moot question.

Apart from this type of technology question being a secondary question in general.

Instead, business leaders need to answer three strategic questions.

The first one is the question of what is most critical for your business. Is it the ERP, CRM, the productivity suite, perhaps ITSM, or something else?

The second one is the one for the vendor that sits squarely in this critical area. This vendor is a good default candidate for getting charged with the cross-application orchestration as well. After all, a considerable number of business processes already touch applications of this vendor.

Analyse this vendor for requirements fit. Does this vendor meet your current and upcoming orchestration needs, and what is the further roadmap? Do you feel comfortable collaborating with this vendor more deeply?

Unless …

And that brings us to the third important question. Do you want a single vendor take care of your main applications and the orchestration of all applications? Or do you prefer to have a second vendor — or multiple — take care of workflow and/or AI orchestration?

In this case, or if you do not feel comfortable about the “default” candidate, do the same analysis with the other vendors that you already work with and perhaps with a select few additional ones that you can identify yourself or with the help of a trusted advisor.

Answering these questions is mission-critical, because choosing your orchestration vendor is now akin to choosing your enterprise AI architecture.

And spoiler alert: Selecting the right partner for cross-application workflow and AI orchestration is even more of a marriage than the selection of business application vendors. Whoever controls your workflows and processes will eventually control your AI strategy. This vendor will be deeply ingrained in the fabric of your business. Pick accordingly and wisely, because a divorce will be very costly, if it is possible at all.

If you do not feel comfortable doing this alone, seek help!

What is your AI orchestration strategy?