In This Guide
⚡ Key Takeaways
- SAP AI is not free in RISE or GROW — baseline AI access is bundled, but meaningful usage requires additional BTP consumption credits that SAP prices at a significant premium.
- Joule entitlements are tiered — the version of Joule you can access depends on your RISE subscription tier, and SAP's sales team will not proactively explain the difference.
- AI FND (AI Foundation) is consumption-billed — every inference call, every embedding, every vector search costs against your BTP credit balance, and those credits deplete faster than most enterprises expect.
- SAP is actively repositioning AI as a separate product line — expect future renewal negotiations to include AI upsells that were previously included in base pricing.
- Benchmark before you sign — SAP's AI unit pricing can vary by 40–60% depending on contract structure, negotiation leverage, and whether you have an independent advisor in the room.
- Usage rights in the Order Form govern everything — the marketing materials say "AI included"; the Order Form defines what that actually means in commercial terms.
What SAP AI Actually Costs — and What They Don't Tell You
SAP AI negotiation tactics are, in 2026, among the most contested territory in enterprise software licensing. SAP's commercial teams have a mission: translate SAP's AI investment into revenue growth. Every RISE with SAP renewal, every S/4HANA Cloud expansion, every BTP conversation is now structured around AI. The CIO hears "AI-powered ERP." The CFO gets an invoice with line items they didn't budget for. Understanding what SAP AI actually costs — and where the pricing levers are — is the first step to negotiating from a position of knowledge.
Let's start with what most enterprises get wrong. When SAP says "AI is included in RISE with SAP," they mean a very specific subset of AI capabilities is included at a baseline level. Generative AI features powered by Joule, deep AI workflows using SAP AI Foundation (AI FND), custom model training on AI Core — these are consumption-billed against your BTP (Business Technology Platform) credit balance. If your BTP allocation is too small, or if you haven't structured the AI consumption rights correctly in your Order Form, you will hit a wall the moment you try to operationalise SAP AI at scale.
The average RISE with SAP contract includes between 50,000 and 150,000 BTP capacity units — a number that sounds large until you start running AI workloads. A single generative AI scenario in SAP S/4HANA using Joule, backed by AI FND, can consume thousands of capacity units per day. Enterprises that signed RISE deals in 2022 or 2023 with AI as a "nice to have" are now discovering that meaningful AI deployment requires a contract amendment — and SAP knows exactly how to price that amendment to maximise its own revenue.
Independent advisory note: Our team has reviewed more than 30 SAP AI contract amendments in the past 18 months. The average increase in BTP spend to support real-world AI deployment was 65% above the base RISE allocation. In 14 of those cases, the initial SAP commercial proposal was 40% above market benchmark. Independent negotiation reduced the final spend by an average of 28%.
The SAP AI Product Landscape in 2026
Before you can apply effective SAP AI negotiation tactics, you need to understand exactly what you are buying. SAP's AI portfolio has grown substantially and the naming conventions are deliberately opaque. Here is the commercial map enterprise buyers need.
SAP Business AI — The Umbrella Brand
SAP Business AI is SAP's marketing umbrella for all AI capabilities embedded across its product portfolio. It covers predictive analytics in SAP Analytics Cloud (SAC), AI-generated content in SuccessFactors, procurement intelligence in Ariba, and the generative AI copilot Joule. From a licensing perspective, "SAP Business AI" is not a line item — it is a framework. The commercial substance sits in the specific products and BTP services underneath it.
Joule — SAP's Generative AI Copilot
Joule is SAP's answer to Microsoft Copilot. It is a generative AI assistant embedded across SAP S/4HANA Cloud, SuccessFactors, Ariba, and other SAP cloud products. Joule processes natural language queries, generates draft documents, summarises financial data, and automates repetitive tasks. From a licensing perspective, Joule access is tiered by your RISE subscription level. Standard RISE customers receive limited Joule entitlements. Additional Joule scenarios require either a tier upgrade or supplemental BTP consumption. See our dedicated guide on what enterprises need to know about SAP AI for the Joule entitlement matrix.
SAP AI Foundation (AI FND)
AI Foundation is the BTP-based infrastructure layer that powers most of SAP's AI capabilities. It includes AI Core (model training and inference), AI Launchpad (management interface), Document Information Extraction, and the vector database services that underpin semantic search and RAG (retrieval-augmented generation) scenarios. AI FND is consumption-billed. Every API call, every document processed, every model inference depletes your BTP capacity balance. This is where the most significant budget surprises occur.
Embedded AI in SAP Applications
Beyond Joule and AI FND, SAP has embedded AI features into specific application modules: cash flow forecasting in SAP S/4HANA Finance, demand sensing in SAP IBP, candidate matching in SuccessFactors Recruiting, spend analytics in SAP Ariba. Many of these features are included in the application licence without additional AI-specific charges. However, the boundary between "included" and "additional consumption" is not always clearly marked in the Order Form, and SAP's commercial teams are skilled at positioning newly included features as new upsells at renewal time.
SAP AI Licensing Models Explained
Understanding the mechanics of SAP AI licensing is central to building effective SAP AI negotiation tactics. There are three primary models to understand:
1. Bundled AI Entitlements
Certain AI features are included in the base RISE or GROW subscription — but "included" has a specific commercial meaning. It means you have access to the feature, not unlimited use of it. Bundled AI entitlements typically come with usage thresholds: a maximum number of AI scenarios activated, a maximum monthly inference volume, or a maximum number of documents processed. When you exceed those thresholds, you move into overage billing — at rates that SAP's Order Form sets unilaterally unless you negotiate them upfront.
2. Consumption-Billed Services (BTP Capacity Units)
The majority of SAP's advanced AI capabilities are consumption-billed through BTP capacity units. These units are the currency of the SAP BTP economy. A capacity unit is a standardised measure of compute, storage, or service consumption. SAP publishes a rate card that maps specific AI services to capacity unit consumption rates — but that rate card is not what most enterprises actually pay. Negotiated rates, volume discounts, and bundle pricing can significantly reduce the effective per-unit cost. Enterprises that sign contracts without benchmarking these rates typically overpay by 30–50% compared to market.
To understand how to track and control consumption costs, read our guide on SAP AI consumption tracking, which covers the monitoring tools and governance frameworks enterprises use to prevent budget overruns.
3. Add-On AI Licence Packages
For specific high-value AI scenarios — SAP's AI-driven business process automation, the Signavio Process Intelligence AI layer, or SAP Analytics Cloud's augmented analytics — SAP offers discrete add-on licence packages. These are priced on a per-user or per-tenant basis and are sold through the standard Order Form process. The strategic risk here is that SAP tends to price these add-ons based on your current user count, then include annual escalation clauses. An 8% annual escalation on an AI add-on that costs £2M/year adds up to £1.8M in extra spend over a five-year term — money that a negotiated fixed-price term would have protected.
Need independent analysis of your SAP AI contract?
Our SAP contract negotiation team has reviewed over 50 AI-related contract amendments and RISE expansions in the past 18 months. We benchmark every SAP AI line item against current market rates and structure agreements that protect your budget across multi-year terms.
Book a Free AI Contract ReviewSAP AI Negotiation Tactics That Work
The following SAP AI negotiation tactics are drawn from direct experience in enterprise negotiations. They are not theoretical — they are the specific moves that shift the commercial outcome in the buyer's favour.
Tactic 1: Separate AI Entitlements from BTP Consumption in the Order Form
SAP's default approach is to bundle AI entitlements within your overall BTP capacity allocation, without specifying per-service consumption rates. This is commercially dangerous. If you exhaust your BTP capacity on non-AI workloads, you lose AI capability without warning. If you add AI workloads unexpectedly, you burn through BTP credits and face an unplanned upsell. Force SAP to specify AI FND capacity as a separate, ring-fenced line in your Order Form with defined consumption rates for each AI service you intend to use. This creates accountability and prevents SAP from using capacity ambiguity as a negotiation lever at your next renewal.
Tactic 2: Demand Joule Entitlement Transparency Before Signing
Joule's entitlement structure is deliberately complex. Different Joule capabilities are available at different RISE subscription tiers, and the specific scenarios available at each tier are defined in a supplemental document — SAP's Joule Feature Matrix — that is rarely shared proactively by SAP's commercial team. Before signing any RISE contract or renewal, demand a complete mapping of which Joule capabilities you receive at your contracted tier, which require additional BTP consumption, and which require a tier upgrade. Get this in writing as an exhibit to your Order Form, not as a marketing presentation.
Tactic 3: Negotiate Consumption Rate Floors, Not Just Caps
Most enterprises focus on negotiating consumption caps — a maximum spend on AI services. That is the wrong focus. What matters more is the consumption rate — the cost per unit of AI service delivery. SAP's standard BTP rate card for AI FND services can be discounted by 20–40% in competitive negotiation situations. Lock in your discounted per-unit rates in the Order Form for the full contract term, with a defined escalation cap (no more than 3% annually). This protects you against SAP repricing AI services mid-term as AI demand grows and SAP gains more market power.
Tactic 4: Include AI in Your ELA (Enterprise Licence Agreement) Scope
If you have an SAP Enterprise Licence Agreement or are negotiating one, push hard to include AI FND capacity and Joule entitlements within the ELA scope. ELAs typically offer the most favourable unit economics because they represent a large, multi-year revenue commitment. SAP is often willing to include emerging product categories within ELA scope when it secures a long-term commitment — the buyer gets better unit economics, SAP locks in a multi-year revenue stream. For guidance on ELA structuring, our SAP ELA advisory service has helped enterprises include AI provisions that saved an average of £2.4M per five-year term.
Tactic 5: Use Competitive Alternatives as Leverage
SAP is aware that enterprises can access comparable AI capabilities through Microsoft Azure OpenAI, Amazon Bedrock, Google Vertex AI, and third-party SAP-compatible AI tools. That awareness creates negotiation leverage — but only if you articulate it credibly. Before your SAP AI negotiation, prepare a concise competitive analysis: the cost of delivering equivalent AI functionality using non-SAP infrastructure, the integration complexity, and the timeline to deploy. You do not need to actually intend to use these alternatives. You need SAP's commercial team to believe that you have done the analysis and that the alternative is commercially viable. That belief alone typically yields 15–25% discount movement.
Tactic 6: Require Pilot Periods Before Full Commercial Commitment
SAP AI capabilities are evolving rapidly. A Joule scenario that is commercially available today may be fundamentally different in 18 months — better, worse, or replaced by a different product. Before committing to a multi-year AI consumption agreement, negotiate a defined pilot period: 90 to 180 days at reduced commitment, with a contractual right to reprice the full-term agreement based on actual pilot consumption data. SAP will resist this — it reduces their certainty of revenue. But a well-structured pilot clause is commercially reasonable and protects you from overpaying for AI capability that does not deliver the expected value.
BTP Credits and AI Consumption: What You're Really Buying
SAP Business Technology Platform (BTP) is the commercial substrate through which almost all SAP AI capabilities are delivered. Understanding how BTP credits work — and how they are consumed by AI workloads — is essential for any enterprise managing SAP AI costs. Our detailed analysis of SAP AI pricing and budget planning covers this in depth, but here are the key principles.
BTP capacity units are purchased in advance, either as part of a RISE or GROW subscription or as a standalone BTP contract. SAP publishes a rate card that specifies the capacity unit consumption for each BTP service. For AI FND, the published rates are: approximately 0.015 capacity units per AI Core inference call, 0.001 capacity units per document page processed by Document Information Extraction, and variable rates for vector database operations depending on collection size and query volume.
These rates sound small in isolation. At scale, they add up rapidly. An enterprise with 10,000 users running Joule-assisted workflows that each trigger 50 AI calls per day generates 5 million inference calls daily — consuming approximately 75,000 BTP capacity units per day, or over 27 million capacity units annually. At SAP's standard published rate, that is a significant annual spend that most enterprises do not budget for when they sign their initial RISE agreement.
The critical SAP AI negotiation tactic here is to conduct a pre-contract AI consumption model before you finalise your BTP commitment. Map the specific AI scenarios you intend to deploy, estimate the usage volume based on user counts and workflow frequency, calculate the expected capacity unit consumption, and then negotiate your BTP capacity allocation and unit price from a data-driven position. Enterprises that do this analysis before signing typically purchase 30–40% less BTP capacity than SAP's commercial team recommends — because SAP's recommendations are calibrated to generate upsell headroom, not to match actual usage needs.
Joule Licensing Deep Dive: What Your RISE Tier Actually Buys You
Joule is SAP's flagship generative AI product and the centrepiece of SAP's AI marketing narrative. Understanding its licensing structure is critical for any enterprise in a RISE renewal or initial RISE negotiation. This is also covered in our sub-page on SAP AI negotiation approach from a commercial strategy perspective.
Joule is not a standalone product you purchase separately. It is a capability embedded within the SAP cloud application suite. The Joule scenarios available to you depend on which SAP cloud applications you have licensed and at which subscription tier. For SAP S/4HANA Cloud Public Edition, Joule scenarios are defined by module: Finance, Procurement, Supply Chain, HR. Not all scenarios are available at the Standard edition — some require the Premium or Advanced edition tiers.
SAP's commercial teams will not proactively present the Joule Feature Matrix during a sales engagement. It exists — it is a detailed technical document mapping specific Joule scenarios to subscription tiers and BTP consumption requirements — but it is typically shared only when a buyer explicitly requests it, and even then, it may be presented in a form that obscures the consumption implications. Demand the full Joule Feature Matrix as part of any RISE negotiation. Read it carefully. Identify the scenarios your business intends to use, confirm they are included at your contracted tier, and document the BTP consumption rate for each scenario in the Order Form.
One critical negotiation point: SAP is adding new Joule scenarios continuously. Your contract should include language that entitles you to new Joule scenarios released during your contract term at no additional charge — provided those scenarios run on your contracted BTP capacity allocation. Without this language, SAP can argue that new Joule scenarios released after contract signing are "new capabilities" that require a separate commercial agreement. We have seen SAP take exactly this position with customers who signed RISE contracts in 2022 and 2023.
Common SAP AI Contract Traps to Avoid
Our review of SAP AI contracts across dozens of enterprise clients has identified six recurring traps that disproportionately expose buyers to unexpected cost. Each is avoidable with the right contract language.
Trap 1: Undefined "Fair Use" Thresholds
SAP's standard contract language for bundled AI features often includes "fair use" or "reasonable use" thresholds that are not numerically defined. This gives SAP the contractual right to notify you that your usage exceeds fair use — and to present a commercial proposal for the "excess." Require numerical definitions: a specific maximum monthly inference count, a specific maximum monthly document volume, a specific maximum concurrent model training jobs. Anything undefined is a future invoice.
Trap 2: AI Overage Rates That Reset Annually
BTP overage rates — the price per capacity unit when you exceed your contracted allocation — are typically defined in your Order Form. What is often not defined is whether your negotiated overage rate applies throughout the contract term or whether SAP can reset it annually to the current list price. This is a critical distinction. Require your Order Form to specify that overage rates are fixed for the duration of the contract term at the same discount percentage as your base BTP rate, not at future list prices.
Trap 3: AI Capabilities Scoped to Current Product Versions Only
SAP's cloud products release new versions on a defined schedule. When SAP introduces a new AI capability in a product upgrade, the default contract position is that this new capability is available to you as part of your subscription — but if it requires incremental BTP consumption, that consumption is billable against your balance. Without proactive monitoring (see our consumption tracking guide), new AI features can activate automatically during a product upgrade and begin consuming your BTP balance without explicit opt-in.
Trap 4: Indirect Access via AI-Generated Outputs
This is an emerging compliance risk. If your enterprise uses SAP AI outputs — documents generated by Joule, data extracted by Document Information Extraction, AI-scored workflow recommendations — to trigger actions in non-SAP systems or by non-SAP users, SAP's indirect access rules may apply. The framework for how SAP charges for AI-mediated indirect access is still developing, but enterprises that have structured their AI workflows without considering indirect access implications may face audit exposure. For independent advice on managing this risk, our SAP indirect access advisory service provides detailed pre-deployment risk assessments.
Trap 5: Vendor Lock-In Through AI Data Residency
SAP's AI services — particularly AI Core and the vector database underpinning Joule — create data residency within SAP's BTP infrastructure. Your fine-tuned models, your proprietary training data, your vector embeddings are stored in SAP infrastructure. This creates a switching cost that SAP's commercial team knows about, even if they do not discuss it openly. Before you invest significantly in SAP AI customisation, understand the data portability provisions in your contract: what you can export, in what format, and what your contractual rights are if you decide to move AI workloads to a different platform.
Trap 6: List Price Escalation on AI Renewals
SAP's AI products are on a steeper price appreciation trajectory than its traditional ERP products. SAP's commercial rationale is that AI capabilities are improving and the value delivered is increasing. That may be true — but it does not mean you should accept SAP's unilateral pricing authority over AI renewals. Negotiate a defined maximum annual price increase for all AI-related line items in your contract: no more than CPI + 2%, with a cap of 5% annually. Without this language, SAP has the right to reprice AI capabilities at renewal at whatever the then-current list price is — and list prices for AI are rising.
See how enterprises have cut SAP AI costs by 30%+
Our SAP licensing case studies include recent engagements where independent advisory helped enterprises restructure AI contract terms, reduce BTP commitment by eliminating over-provisioning, and introduce consumption rate protections that saved millions across multi-year terms. Book a free consultation with our SAP AI licensing team.
Book Free Consultation →Deep-Dive Guides in This Series
Frequently Asked Questions
Is Joule included in all RISE with SAP contracts?
Joule access is included in RISE with SAP subscriptions, but the scope of that access depends on your subscription tier and which SAP applications you have licensed. Basic Joule scenarios for core ERP processes are generally included. Advanced Joule capabilities — particularly generative AI features that require significant BTP compute — may require either a tier upgrade or additional BTP capacity purchase. Always request the full Joule Feature Matrix before signing.
How do I calculate how many BTP capacity units I need for SAP AI?
Start with the specific AI scenarios you plan to deploy. For each scenario, identify the underlying AI FND services involved (AI Core inference, Document Information Extraction, vector search, etc.). Estimate the monthly usage volume based on user counts and workflow frequency. Apply SAP's published consumption rates to calculate expected monthly capacity unit consumption. Add a 30% buffer for peak usage and unexpected workloads. Compare the total against your contracted BTP allocation. This analysis should be done before contract signing — not after you start deploying AI and discover a gap.
Can SAP change the BTP consumption rates for AI services mid-contract?
Under SAP's standard contract terms, SAP can adjust BTP rate cards with notice — typically 90 days. This is one of the most important negotiation points for any enterprise deploying AI at scale: lock in your AI service consumption rates for the full contract term, not just for the initial period. This requires explicit language in your Order Form and may require negotiation effort, but it is commercially achievable. We have successfully locked in fixed AI consumption rates for multi-year RISE and BTP contracts.
What leverage do enterprises have when negotiating SAP AI pricing?
You have more leverage than you think. First, competitive alternatives to SAP AI (Azure OpenAI, Amazon Bedrock, third-party tools) provide credible benchmarks. Second, your overall SAP spend — particularly if you have a large RISE or GROW contract — creates volume leverage. Third, the timing of your negotiation matters: SAP quarter-end and year-end create urgency that translates to commercial flexibility. Fourth, an independent advisor with market benchmark data substantially changes the dynamic — SAP knows that advisors who have reviewed multiple contracts will push back on above-market pricing.
Should SAP AI be included in our ELA scope?
For enterprises spending more than £5M annually with SAP, including AI capabilities in an Enterprise Licence Agreement is worth analysing carefully. ELAs offer the most favourable unit economics but also lock in a multi-year commitment. The right answer depends on your confidence in your AI deployment roadmap. If you are committed to scaling SAP AI across multiple applications over 3–5 years, ELA inclusion typically delivers the best unit economics. If your AI deployment plans are uncertain, a consumption-based arrangement with negotiated rate protections may be more appropriate.
📬 SAP Licensing Intelligence
Independent SAP AI Insights — Free
Expert analysis on SAP AI licensing, BTP consumption management, and contract negotiation. No vendor affiliation.
Independent SAP licensing advisory — not affiliated with SAP SE. SAP, S/4HANA, RISE with SAP, Joule, BTP, and all SAP product names are trademarks of SAP SE. Our advice is 100% buyer-side.