The Future of CRM: How AI, Composable Architecture, and Real-Time Data Are Reshaping SAP CX

CRM is going through its biggest architectural shift since the move to cloud. I see three forces driving it. AI agents that can handle sales and service tasks on their own. Composable architecture that lets you swap out components instead of buying monolithic suites. And real-time data unification that makes personalisation possible at scale instead of as a marketing promise.

Here’s what each of those means in practice and where SAP CX fits.

TL;DR: By 2028, AI agents will outnumber sellers by 10x, yet fewer than 40% of sellers will report improved productivity (Gartner, 2025). The gap between AI availability and AI productivity means the winners won’t be who adopts AI first, but who integrates it into real workflows. SAP CX is embedding Joule AI across Sales, Service, and Commerce while shifting to composable architecture and real-time CDP as the data backbone.

How Is AI Changing CRM?

Sixty-seven per cent of SAP’s Q4 2025 cloud orders included business AI, up over 20 points from Q3 (CX Today, 2026). AI adoption in CRM is accelerating. That’s obvious. But what does “AI in CRM” actually mean once you cut through the noise?

Joule AI in Sales Cloud V2. Joule acts as a copilot for sales reps: summarising accounts, recommending next actions, drafting emails, scoring opportunities based on deal patterns. Adoption grew ninefold over 2025. What makes it work is that Joule sits inside the sales rep’s existing workflow. No context-switching to a separate tool. That matters more than people think.

AI-powered service routing. Service Cloud V2 uses AI to categorise incoming cases, assign them to the right agent based on skills and workload, suggest knowledge base articles. Cuts average handling time, improves first-contact resolution. Solid, practical stuff.

Intelligent Selling Services in Commerce. Commerce Cloud’s ISS uses machine learning for product recommendations, adaptive search, personalised merchandising. It analyses browsing and purchase behaviour to surface the most relevant products. Not flashy, but it moves conversion rates.

Predictive analytics in Emarsys. AI-driven customer segmentation, churn prediction, campaign optimisation. Emarsys predicts who’s likely to buy, churn, or re-engage, and automates the marketing response. The measurement is built in, which is nice.

Here’s the counter-trend worth watching though: by 2030, 75% of B2B buyers will prefer sales experiences that prioritise human interaction over AI (Gartner, 2025). AI should augment human relationships. Not replace them. I reckon most of us in the CX space already know this intuitively, but it’s good to see the data backing it up.

What Does Composable CRM Architecture Mean?

Worldwide public cloud spending is forecast at $723.4 billion in 2025, up 21.5% from 2024 (Gartner, 2024). All that cloud adoption enables a shift from monolithic CRM suites to composable architectures.

Composable CRM means you pick best-of-breed components for each function and connect them through APIs. Not buying an all-in-one suite and living with whatever each module offers. Want SAP Sales Cloud V2 for sales but Emarsys for marketing and a different CDP? Composable architecture makes that possible without everything falling apart.

SAP CX is designed for this. Each product exposes clean REST APIs. Extensions run on BTP as independent services. Integration Suite connects everything. You’re not locked into using all five products. Mix SAP CX products with third-party tools where they fit better.

The Clean Core principle reinforces this: keep each product standard, extend through APIs, don’t let custom code create monolithic lock-in. It’s a proper architectural philosophy. Not a label.

How Is Real-Time Customer Data Changing CX?

The global CRM market is valued at $112.91 billion in 2025 (Fortune Business Insights, 2025). A growing chunk of that investment goes toward real-time data infrastructure. For good reason.

Traditional CRM systems sync data in batches. Nightly. Hourly if you’re lucky. Real-time CRM means events flow between systems as they happen. Customer opens a support ticket, the sales rep sees it immediately. Customer makes a purchase on the storefront, their profile updates across all systems in real time.

That changes everything about how you serve people.

SAP enables this through SAP Event Mesh, an event broker on BTP that distributes events across CX products and extensions. Something happens in one system, interested services get the notification instantly.

Real-time data enables things batch processing simply can’t do: personalised web experiences based on the last 5 minutes of browsing, dynamic pricing based on current inventory, proactive service outreach when a sensor detects an issue, sales alerts when a key account’s behaviour shifts. Prima vista it sounds like marketing hype, but we’ve built these patterns for clients and they work.

What Should You Do About These Trends?

Forty-three per cent of organisations say generative AI influenced their ERP decisions in 2025, up from 14% in 2023 (ERP Today, 2025). AI has moved from nice-to-have to de facto decision factor. So here’s what I’d actually do about it.

Short term (next 12 months). Turn on Joule AI in your existing SAP CX products. You don’t need a new project for this. Just activate the AI features already in your subscription. Start with account summaries and next-action recommendations in Sales Cloud V2. Low effort, immediate value.

Medium term (12-24 months). Invest in real-time data infrastructure. If you’re still running batch integrations (and be honest, most of you are), move to event-driven patterns with SAP Event Mesh. Real-time data is the foundation for AI features that actually deliver. Without it, you’re feeding stale data to smart algorithms. That’s a waste.

Long term (24-36 months). Evaluate your architecture for composability. Are you locked into components that aren’t pulling their weight? Could you swap a weak module for a best-of-breed alternative? Clean Core architecture makes this possible without rebuilding everything.

FAQ

Will AI replace sales reps?

No. By 2030, 75% of B2B buyers will prefer human interaction (Gartner, 2025). AI will handle repetitive tasks (data entry, scheduling, research) while reps focus on relationships, negotiations, and complex deals. The human touch still matters. A lot.

What’s the biggest risk of adopting AI in CRM too fast?

Over-reliance on AI recommendations without human judgement. AI models are only as good as their training data. If your CRM data is incomplete or biased, the recommendations will be too. Start with AI as a suggestion engine, not a decision engine.

Is composable CRM more expensive than a suite?

Depends. Composable can reduce licensing costs (pay only for what you actually use) but may increase integration costs (connecting multiple tools). For mid-market companies, starting with SAP CX and selectively replacing components is typically more cost-effective than building composable from scratch. Fair enough on both approaches.

How does SAP CX compare to Salesforce for AI capabilities?

Both invest heavily. Salesforce has Einstein and Agentforce; SAP has Joule. SAP’s edge: Joule is embedded directly in CX workflows and can access ERP data natively (real-time pricing, inventory, order history). Salesforce’s edge: larger ISV ecosystem and more third-party AI integrations. Different strengths for different situations.

When will AI agents handle full sales cycles autonomously?

Not soon. AI agents can handle individual tasks (qualifying leads, drafting emails, scheduling meetings) but lack the judgement for complex B2B negotiations. Expect AI to handle 60-70% of repetitive sales tasks by 2028, with humans managing the relationship-critical stuff. And I reckon that’s where we want the line to be.