News from SAP Connect 2026: the Autonomous Enterprise has Arrived

Key Takeaways

• SAP assistants now carry out defined workflow steps instead of only answering questions, while people keep oversight.
• Six new or enhanced CX assistants cover campaigns, content, B2B orders, quotes, sales compensation, and service planning.
• SAP now maintains the connections between its cloud ERP and other SAP apps, so you skip separate integration projects.
• Most CX assistants reach general availability in Q1 2027, which gives you about one quarter to get your data ready since the assistants only perform as well as the data underneath them.

SAP Connect ran from 5 to 7 October 2026. On 6 October, SAP used its global keynote to show customers such as Nestlé, Novartis, and Morgan Foods running AI assistants inside real business processes. It also announced a new wave of assistants for SAP Customer Experience – the software behind your marketing, commerce, sales, and service teams – and for cloud ERP, the system that runs finance, orders, and inventory behind the scenes.

Our Chief Customer Officer Eric Schreiner was on the ground in Las Vegas. Below, we share what we saw, what we think it means for your business, and what still needs proving.

What SAP Announced at Connect: Assistants That Carry Out Work

Until now, most business AI answered questions. At SAP Connect, SAP’s message was that its assistants now do work: they prepare and carry out defined steps inside your workflows, using the same data, permissions, and audit trails your teams already rely on. SAP CEO Christian Klein put it plainly in the press release: “The Autonomous Enterprise has arrived.”

Two details matter more than the headline. First, customers set the level of autonomy – so you can start with an assistant that prepares work for approval and give it more freedom as trust grows. Second, SAP pairs the assistants with governance, including a central inventory of every AI agent in your company and a setup certified to ISO/IEC 42001, the international standard for managing AI.

The early results are modest but concrete. SAP reports around 20% productivity gains across finance, HR, and procurement among the 110,000 SAP employees who already use Joule Work, its new plain-language interface.

The assistants also connect to third-party AI agents through the open Agent2Agent protocol. That matters if your landscape includes tools beyond SAP, such as Google Cloud.

For you, the realistic first payoff is time: your teams spend less of the day retyping, chasing, and checking, and more on exceptions and customers. How much you gain depends on the level of autonomy you choose. An assistant that drafts a quote or handles a commission dispute only helps if you decide upfront what it may do alone and what goes to a person. SAP lets you choose that limit, and choosing it is your job.

Eric makes a good point about scaling autonomy: “Automation inside SAP processes is not new – we have had rules, workflows, and scoring models for years. What is new is an assistant that works across several systems and can explain what it did. That makes governance the real project: who approves, what gets logged, and what happens at the exception. Get that right and autonomy scales.”

We have already covered the SAP Autonomous Enterprise vision and the SAP Business AI Platform underneath it. Here, we focus on what is new and what it means for CX and cloud ERP.

Autonomous Customer Experience: New Generation of AI Assistants

Building on the Autonomous CX area introduced at SAP Sapphire in May 2026, SAP added six assistants to its CX portfolio – three enhanced and three new. General availability is planned for Q1 2027.

Here is what each assistant does, according to SAP, and what it could mean for you:

• Campaign Assistant (enhanced): activates omnichannel campaigns based on customer and business signals, so campaigns can react to things like stock levels or customer behavior instead of waiting for the next planning cycle.
• Content Assistant (enhanced): helps marketers assemble personalized, channel-ready content and adapt it for each channel, which means fewer hours spent rewriting the same message for every channel and region.
• Shopping Assistant (enhanced): adds B2B email order processing by checking customer and product details in incoming orders, so orders stop being retyped by hand and your team handles only the exceptions.
• Quoting Assistant (new): helps sales teams create accurate, compliant quotes and customer-ready proposals in one workflow, which can mean shorter deal cycles and fewer pricing mistakes that quietly erode margin.
• Incentive Compensation Assistant (new): lets salespeople raise compensation disputes and shows how crediting and payments were calculated, so sales and finance spend less time arguing over commissions.
• Service Planning and Operations Assistant (new): connects commercial planning, logistics, post-service activities, and performance data from first contact to final invoice, so fewer steps get lost between teams and invoices go out faster.

An assistant that works across teams is only as good as the information it can reach, and SAP says its assistants draw on shared business context such as customer interactions, orders, contracts, pricing, and payment history. In most enterprises, that context sits in a CX system, a cloud ERP, and a few spreadsheets nobody mentions in steering meetings.

We see the same logic in our own projects. The AI beauty advisor we built with DOUGLAS works because it reads live product data and each shopper’s profile. The new SAP assistants apply that principle across marketing, sales, and service.

Two cautions before you plan around these assistants. Q1 2027 is a planned date, and SAP itself flags such statements as subject to change, so budget for pilots in 2027 rather than full rollouts. And data quality still decides the result: analysts quoted by CIO.com note that the closer agents get to core workflows, the more data integrity, privacy, and accuracy matter.

“In the CX portfolio, the hard part has always been the hand-off to the back office: a quote that does not reflect real pricing, an order that cannot see stock. These assistants target exactly those seams. So the question for SAP is not what an assistant can generate, but which systems it can read from and write to on day one.” says Eric.

Cloud ERP: Managed Integrations, Payments, and Data

If the CX assistants are the visible part, the cloud ERP news is what makes them work. Four announcements stand out.

SAP now maintains the connections between its apps

In the latest SAP-managed public cloud ERP release, SAP delivers, maintains, and upgrades the integrations to apps such as SAP Ariba (procurement), SAP Sales Cloud, SAP Subscription Billing, and SAP SuccessFactors (HR) as part of your cloud ERP subscription.

You skip the separate integration projects, upgrade work, and heavy regression testing that normally come with them. The Ariba and Sales Cloud connections are available now, Subscription Billing follows in November 2026, and SuccessFactors in Q4 2026.

One order layer across several ERPs

SAP extended its Order Management Services so companies can coordinate B2B orders, inventory, pricing, and fulfillment across SAP S/4HANA and other ERP systems, without consolidating them first.

New Joule agents help operations teams spot and respond to disruptions with less manual effort. General availability is planned for Q1 2027.

Payments inside the ERP

SAP Pay executes a payment and reconciles it against the invoice or purchase order as soon as the invoice is approved, so finance teams skip bank file handling and manual matching. It is available now in the United States and the United Kingdom only, so European customers should ask SAP about timing.

Cleaner data for the assistants

SAP Relio, part of SAP Business Data Cloud, merges the scattered records for the same customer, supplier, or product across SAP and third-party systems into one trusted version. It is available now, and it adds to the data investments we covered when SAP announced its acquisitions of Dremio and Prior Labs.

The common thread is a shift in where your effort goes. Less of it goes into building and testing connections between SAP apps, and more goes into preparing the ERP those apps connect to. SAP-managed integrations are built around SAP’s standard processes, so an ERP with a decade of custom code may need clean-up before it benefits fully. That is why we launched our ERP practice in July: to connect the front office to the back office without carrying years of custom code along.

“An agent is only as useful as what it can see and do. If your storefront, your CRM, and your ERP each hold a different version of the customer and the order, the agent either guesses or stops. Connect the front office to the back office first, so it can pull stock, pricing, and payment history from across the business and act on them.” 

The multi-ERP order layer deserves a second look. Groups that grew through acquisitions often run several ERPs, and a full consolidation can take years. A shared order layer lets them give customers and sales teams one view of orders and stock while that bigger project waits.

However, note that SAP-managed does not mean everything-managed. The new integrations cover SAP’s own apps, so your search, product data, marketplace, and loyalty tools still need to connect to the same data – and that is where we usually see projects slow down. SAP Relio helps reconcile records, but it cannot decide what “customer” means in your business. That is a business decision first and a technology task second.

How to Prepare before Q1 2027

SAP Connect 2026 showed AI assistants doing real work in real companies. The CX and cloud ERP announcements go after the same targets: the hand-offs, retyping, and waiting that slow down revenue work. What you gain will depend less on which assistant you switch on and more on whether your customer data, your ERP, and your decision rules are ready for it.

To get ready for Q1 2027, we suggest you:

• pick two or three workflows where your teams lose the most time between systems, such as quote approvals
• measure how long they take today, so you can prove the improvement later
• check that customer, product, and pricing data is consistent across your CX system and your ERP
• agree which decisions an assistant may take on its own and which need approval

If you want a step-by-step Autonomous Enterprise implementation roadmap, or guidance on the people side of an agentic AI adoption strategy, we have written up both.

Most of the new CX assistants are planned for Q1 2027, so the preparation window is now. To see where you stand, our Agentic AI Readiness Checklist helps you assess your data and digital systems at your own pace. And if you would rather talk it through, get in touch – Eric and the team would be happy to guide you through the process.