All 25 bulletins from Orlando, what’s truly transport vs. what’s advertising — and what it means for the doc and information basis of the autonomous enterprise.
SAP’s Sapphire 2026 in Orlando was probably the most AI-dense keynote within the firm’s historical past. Christian Klein did not unveil a roadmap. He unveiled capabilities — 25 of them, by our depend — starting from a elementary repositioning of SAP itself to an embedded Mannequin Context Protocol server, a €1B+ dedication to tabular basis fashions, and a robotics partnership packing bins in SAP’s personal logistics heart.
And but, whereas Klein was on stage saying the Autonomous Enterprise, SAP’s inventory sat ~28–32% down year-to-date. Analysts stored Purchase scores however quietly reframed AI as “an adoption subject, not a income driver but.”
That pressure — between the boldness of the imaginative and prescient and the persistence of the market — is an important factor to know about Sapphire 2026. Beneath is the entire breakdown: all 25 bulletins numbered as they seem, grouped into the seven themes that matter.
Theme 1: The Repositioning
1. The “Autonomous Enterprise” is the brand new imaginative and prescient. SAP is turning into a “enterprise AI firm,” not only a software program firm. Joule, SAP’s assistant, delivered the framing line from the keynote stage. The implication: SAP is not competing with Workday and Oracle anymore — it is competing with Salesforce, ServiceNow, and the agentic AI gamers for the layer that acts on enterprise information, not simply information it.
2. SAP Enterprise AI Platform unifies SAP Enterprise Expertise Platform (BTP), Enterprise Knowledge Cloud (BDC), and Enterprise AI right into a single stack. It is a consolidation greater than a brand-new product — however the sign issues: SAP is positioning one basis layer above the apps, not three.
Our learn: The repositioning is significant. However SAP has tried variations of this story earlier than. The query is not whether or not SAP needs to be a enterprise AI firm. It is whether or not it could truly ship working AI in manufacturing, the place each prior try has stalled.
Theme 2: The Agent Stack
3. SAP Data Graph — SAP turned 452,000 tables and seven.3M fields into machine-readable semantics for brokers to motive in opposition to. That is the under-the-hood plumbing that lets Joule brokers perceive SAP’s personal information mannequin.
4. 50+ Joule Assistants — domain-specific chat/copilot helpers throughout finance, procurement, provide chain, HR, and buyer expertise.
5. 200+ specialised AI brokers / “Autonomous Suite” — the brokers the assistants orchestrate. Consists of Joule for Builders and an ABAP code conversion agent going GA. The one Joule agent with a tough, publicly-stated metric: a 40% effectivity goal.
6. Joule Studio 2.0 GA — the build-and-govern setting for brokers. Free design-time entry by way of December 31, 2026.
Our learn: The Data Graph is probably the most underappreciated announcement of Sapphire. With out semantic understanding throughout 452K tables, no agent can motive coherently throughout SAP information. That is structural plumbing that compounds for years.
The Joule Assistants are more durable to guage. SAP has been transport assistants for years — S/4HANA copilots, Joule v1 — and nothing has labored in manufacturing at scale. These are UI helpers, not workflow replacements, and that distinction tends to vanish in advertising supplies.
The 200+ agent quantity is spectacular on paper, however Salesforce ran this precise playbook with Agentforce twelve months in the past, and the manufacturing actuality fell nicely in need of the pitch. The 40% effectivity goal on the ABAP agent is the one quantity with enamel. Watch what will get quantified subsequent — and what does not.
Theme 3: The Ecosystem Opens
7. SAP AI Agent Hub — a vendor-agnostic market for brokers (SAP and third-party). GA Q3 2026. SAP has reportedly taken 680+ accomplice submissions.
8. Mannequin Context Protocol (MCP) embedded in Enterprise Knowledge Cloud and Reltio. Third-party brokers — constructed exterior SAP — can plug into SAP’s information layer by way of an open protocol, not a proprietary integration.
9. n8n partnership and SAP fairness stake at a reported $5.2B valuation. Workflow orchestration embedded straight in Joule Studio.
10. Strategic partnerships throughout the stack: Anthropic (Claude turns into a major reasoning mannequin embedded throughout Joule), AWS (zero-copy integration with Athena), Google Cloud + Microsoft (bidirectional Agent-to-Agent / A2A protocol, GA This autumn 2026), Mistral and Cohere (sovereign mannequin choices), NVIDIA (OpenShell because the safe runtime for Joule Studio).
Our learn: That is the structural shift that issues most. For the primary time, SAP is opening outlined interop paths — MCP, A2A, the AI Agent Hub — that allow specialised AI instruments plug into the SAP ecosystem with out proprietary engineering. The walled backyard is turning into a community.
Virtually: if Joule Assistants get a finance or procurement workflow 70–80% of the way in which there, the final 20–30% — correct doc extraction, vendor information enrichment, contract intelligence, edge-case automation — can now be dealt with by specialised brokers speaking over MCP. That is a essentially completely different vendor panorama than the one we had at Sapphire 2025.
Theme 4: Knowledge Basis Strikes
11. Reltio acquisition closed — grasp information administration for SAP and non-SAP environments. Reltio’s power is connecting to non-SAP sources and dealing with advanced multi-source information — precisely the place SAP’s native stack has been skinny.
12. Dremio acquisition (information lakehouse), anticipated to shut Q3 2026.
13. €1B+ dedication to Prior Labs for tabular basis fashions (SAP-RPT-1.5). Related for structured information extraction — turning messy enterprise information into structured rows brokers can devour.
Our learn: Three strikes of this magnitude is SAP saying out loud what enterprise groups have recognized for years: the information flowing into SAP — from paperwork, from non-SAP techniques, from unstructured sources — just isn’t in a state the place brokers can reliably devour it. Reltio addresses cross-system grasp information. Dremio addresses the lakehouse layer. Prior Labs addresses unstructured-to-structured conversion.
What’s notably not addressed is the doc layer itself. Invoices, contracts, POs, vendor onboarding paperwork, receipts. The textual content and structured information inside enterprise paperwork nonetheless has to land in clear, validated type in SAP’s information layer earlier than any agent can act on it. That extraction layer is the place specialised gamers proceed to win — and the place SAP is, once more this 12 months, leaving the sector open.
Theme 5: Buyer Proof Factors
14. JPMorgan is shifting its basic ledger to SAP. The one largest buyer win introduced at Sapphire and, frankly, the 12 months.
15. 26 industries coated at GA with industry-specific brokers — shopper/retail, oil & fuel, mining, engineering, agriculture, banking, insurance coverage, and extra. H&M’s unified commerce went reside as a keynote demo.
16. Bodily AI — Cyberwave robotics partnership. Autonomous packing in SAP’s personal logistics heart.
17. Expanded sovereign cloud, together with India.
Our learn: The JPMorgan GL win is probably the most strategically important buyer announcement of 2026. Banks shifting a basic ledger is uncommon, conservative, and alerts deep institutional dedication. That alone offers the Sapphire narrative extra credibility than the agent depend does.
The 26-industry protection is extra breadth than depth — count on the {industry} brokers to be uneven in manufacturing, with robust ones in retail and provide chain and weaker ones in regulated verticals the place information and compliance complexity is increased.
The Bodily AI demo grabs headlines however will not transfer enterprise budgets in 2026. Sovereign cloud growth to India, alternatively, is significant — information residency has been a real blocker for SAP cloud adoption in Indian enterprises, and eradicating it opens a big market.
Theme 6: The Migration Downside
18. Parloa partnership for AI brokers in SAP Service Cloud.
19. Palantir + Accenture as implementation companions for advanced migration situations.
20. RISE with SAP reset: 3 Joule Assistants contractually included in 12 months one, with the Max Success Plan extending adoption throughout the enterprise.
21. SAP GROW reset: 20+ AI assistants from day one (positioned for smaller and midmarket clients).
22. €100M accomplice fund for companions deploying AI assistants or constructing on Joule Studio.
23. Migration assistants: 35%+ effort discount, as much as 50%, for S/4HANA migrations.
24. AI entry requires cloud migration. Prospects should be on RISE, or transfer not less than 50% of upkeep spend to cloud, to make use of the brand new AI options. Roughly two-thirds of SAP’s base continues to be on ECC or on-prem.
Our learn: That is the place the advertising meets the wall. S/4HANA migration has been a slog for years — most ECC clients are reluctant, and the migration tooling enhancements really feel like a pressured acceleration push. Tying AI entry to cloud migration creates an express forcing perform: SAP needs enterprises emigrate to entry brokers.
However the math is uncomfortable. Two-thirds of the SAP base is on ECC, and most of them do not wish to transfer on SAP’s most popular timeline. SAP has successfully determined to not assist AI for these clients. That creates an actual and rising marketplace for tooling that lets ECC clients undertake AI on the doc and course of layer whereas staying on ECC till they’re able to migrate on their very own phrases.
The Parloa partnership tells you one thing else: SAP is keen to white-flag particular useful areas. Parloa is a customer support AI play, and SAP introduced them in as a result of SAP could not sustain with Salesforce in CS. Count on that sample — herald a specialist the place the hole is widest — to repeat throughout different features.
Theme 7: The Market Response
25. SAP’s inventory was down ~28–32% YTD heading into Sapphire. Analysts maintained Purchase scores however flagged AI as “an adoption subject, not a income driver but.”
The €100M accomplice fund, the AI-density of the keynote, the rapid-fire acquisitions — these are responses to investor skepticism as a lot as they’re responses to buyer demand. SAP wanted Sapphire 2026 to appear like an AI firm. It did. The market hasn’t but been satisfied it is one.
That hole between narrative and conviction is probably the most sincere sign popping out of Orlando.
What This Truly Means for Enterprise AI Patrons
Strip away the keynote choreography and Sapphire 2026 tells a coherent story for SAP clients:
1. The agent layer is actual however immature. Joule Assistants and the 200+ brokers are transport, however they’re finest understood as copilot helpers right now, not autonomous workflow replacements. Salesforce’s Agentforce is the related case research, and that hasn’t but lived as much as its pitch.
2. The interop story adjustments the seller panorama. MCP in BDC and Reltio, A2A by This autumn, the AI Agent Hub market by Q3 — collectively, these bulletins say SAP is opening up. Specialised brokers will plug in. The perfect-of-breed period is again, even inside SAP environments.
3. The info and doc basis drawback is larger than SAP is letting on. Three acquisitions (Reltio, Dremio, Prior Labs) and a €1B+ dedication to tabular basis fashions. SAP is shopping for its manner into an information layer it does not have. The doc layer — invoices, POs, contracts, vendor information — sits adjoining to all of those strikes and stays the dominant bottleneck for agent reliability.
4. Two-thirds of SAP clients are stranded. ECC customers cannot entry the brand new AI options with out migrating. Most will not migrate on SAP’s most popular timeline. There’s an actual, rising, and underserved marketplace for AI tooling that works on ECC right now.
5. The 80% accuracy bar Klein set is a brutal filter. “Eighty % accuracy is not enough when operating mission-critical companies.” Each agent shipped in 2026 will probably be measured in opposition to that line. Most will not clear it with out enterprise-grade structured inputs upstream.
The place Nanonets Matches
The autonomous enterprise has a quiet bottleneck: brokers are solely pretty much as good because the paperwork and structured information they’re grounded in.
Nanonets builds the doc AI layer that turns the unstructured enter flowing into SAP — invoices, POs, contracts, vendor information, receipts — into the structured, audit-ready, enterprise-grade information that Joule Assistants and SAP’s 200+ brokers want to truly work in manufacturing.
We combine with SAP S/4HANA and SAP Ariba right now. We additionally work in ECC environments the place Joule does not but attain — which issues for the two-thirds of the SAP base that will not migrate on SAP’s timeline. And we’re constructed explicitly for the accuracy bar Klein set publicly in Orlando.
With MCP now embedded in SAP Enterprise Knowledge Cloud and Reltio, and the SAP AI Agent Hub going GA in Q3, the combination paths for specialised AI instruments into SAP environments are about to be essentially extra open than they’ve ever been. We’re monitoring these timelines intently and constructing towards them.
Should you’re rethinking your doc and information basis for the autonomous enterprise period — whether or not you are on RISE, GROW, or nonetheless on ECC — we would love to check notes.
Sources: SAP Information Middle, SAP Neighborhood, ERP At this time, Constellation Analysis, SAPinsider, Channel Insider, ad-hoc-news, SAVIC Applied sciences. Quotes and metrics drawn from Christian Klein’s Sapphire 2026 keynote and the official SAP Sapphire 2026 bulletins.
