Salesforce Agentforce
Teams evaluating an AI customer service platform ask the same three questions: can we self-host, can our team own it, and what does it actually cost? Here's how Rasa compares on the dimensions that decide enterprise deals
Two platforms, two opposite philosophies
Salesforce Agentforce offers AI agents within Service Cloud with full CRM context, Einstein GPT for replies and summaries, and native telephony via Service Cloud Voice.
Rasa is an enterprise conversational AI platform built on a self-hosted, developer-owned architecture. Patented dialogue management (CALM) delivers guided governance: business logic controls high-risk actions through explicit policies, regardless of LLM output. Native voice (Twilio, AudioCodes, Genesys), 100% on-prem or private cloud, transparent conversation-volume pricing. Customers include N26, Deutsche Telekom, Helvetia, Autodesk.
One platform for voice and chat, running in your environment
Rasa runs the same guided-governance engine across phone and chat, fully self-hosted. Your team configures flows, policies, and integrations directly, with no managed-service dependency.
- Native voice over Twilio, AudioCodes, and Genesys, sharing context with chat.
- Explicit policies control high-risk actions on regulated workflows, regardless of LLM output.
- Deploy on-prem, in private cloud, or air-gapped, with no customer data leaving your perimeter.

Side-by-side on the six dimensions that decide enterprise deals
Five dimensions, side by side
CRM-Native Context vs. Flexible Multi-Channel Architecture
Salesforce Agentforce delivers AI agents within Service Cloud, giving agents and customers full CRM context in the same interface. Einstein GPT powers generative replies and ticket summaries, and Service Cloud Voice handles native telephony through Amazon Connect. For enterprises already committed to Salesforce, this integration is seamless: no data switching, no separate logins, and every Salesforce object flows into the AI logic. Regional data residency via Hyperforce addresses compliance. However, per-conversation pricing ($2 each) compounds at enterprise volume, licensing is complex across tiers, and you are locked into Salesforce's roadmap. Agentforce suits contact centers where CRM depth is the primary lever, but the lack of self-hosted options disqualifies it for regulated industries requiring on-premise deployment.
Rasa's customer-owned architecture: self-hosted from day one, patented dialogue management for guided governance, full code-level extensibility, native voice via Twilio/AudioCodes/Genesys, and transparent conversation-volume pricing. N26, Deutsche Telekom, and Helvetia run Rasa for regulated workflows.
Deployment Model & Data Sovereignty
Salesforce Agentforce is cloud-only, hosted in Salesforce's data centers or regional Hyperforce instances. Hyperforce allows. Data residency via Hyperforce meets. Multi-tenancy: Agentforce runs on shared Salesforce infrastructure, but data is logically separated. No self-hosted option in current public materials. No on-premise deployment. For regulated industries requiring on-premise deployment, Agentforce is disqualified. Hyperforce mitigates some data residency concerns but still runs Salesforce's infrastructure, not yours. Switching away from Salesforce means re-implementing dialogue logic and retraining models elsewhere—lock-in is significant.
Rasa supports cloud, on-premise, and hybrid deployment out of the box. Your data never leaves your infrastructure unless you choose to send it elsewhere. On-premise deployment is included in the open-source edition at no cost; enterprise support adds dedicated implementation specialists and SLAs. Unlike managed services, Rasa gives your team full control: choose your hosting provider, VPC, Kubernetes distribution, or even air-gapped environments. Compliance teams approve faster because your security team reviews the architecture. No vendor dependency on SaaS uptime. For regulated industries (healthcare, finance, legal), on-premise is non-negotiable. Rasa delivers it without premium surcharge or complex licensing.
Security & Compliance
Salesforce has. Agentforce's compliance posture is strong, but audit depth depends on your Salesforce environment's configuration. For healthcare (HIPAA), Salesforce requires custom BAA and may require Hyperforce in specific regions. For financial services (PCI-DSS, SOX), compliance is shared: Salesforce handles platform, you handle data governance.
Rasa runs on your infrastructure or your chosen cloud provider. You control encryption at rest (your key management), encryption in transit (TLS 1.2+), and access controls (IAM). Rasa does not hold customer data in a shared multi-tenant system. Enterprise Rasa includes SOC 2 Type II compliance, GDPR readiness, and audit-ready logging via OpenTelemetry. For HIPAA, you implement HIPAA-compliant infrastructure; Rasa is platform-agnostic. For PCI-DSS, on-premise deployment gives your compliance team full visibility. Rasa's architecture is inspectable: every webhook, every API call, every data flow is traceable. This transparency enables fast audit cycles and regulatory approval.
Developer Experience & Integration
Agentforce is designed for Salesforce admins and developers. Setup requires Salesforce admin knowledge (user provisioning, workflows, custom fields). Configuration via Salesforce UI; development via Apex, Flow, or declarative tools. For Salesforce-experienced teams, Agentforce setup is fast. For teams new to Salesforce, the learning curve is steep. Switching costs are high: moving to Rasa or another platform requires re-implementing Apex and re-training models.
Rasa is built for engineers. Start in VS Code or your IDE of choice; deploy via GitHub Actions or Jenkins. Rasa provides SDKs (Python, JavaScript), OpenAPI specs, and REST APIs. Extend via custom Action Servers (Python, JavaScript, Go). Integrate any CRM, database, or third-party service via REST or MCP. No low-code UI required; if your team prefers code, Rasa is fully scriptable. Rasa Playground and Rasa Studio provide visual debugging for non-engineers. For large teams, Rasa's dialogue state is event-based and queryable—understand exactly why the AI made a decision. CI/CD integration is native. Deployment is reproducible: commit your models, test in CI, deploy to production via container orchestration.
Pricing & ROI
Salesforce Agentforce pricing has multiple components: (1) Agentforce license: $2/conversation, (2) Service Cloud license: $25–$165/user/month, (3) Total cost compounds quickly. A deployment with 1M conversations/month = $2M/year for Agentforce alone, plus Service Cloud seats. For contact centers with 100+ agents, annual cost easily exceeds $500k. ROI calculation is opaque due to layered pricing. Per-conversation model means cost scales linearly with volume, unlike Rasa's fixed annual licensing.
Rasa Developer Edition is free, forever. One bot per company; up to 1,000 external conversations per month. Community support via GitHub and Rasa Forum. Rasa Enterprise is transparent annual licensing based on conversation volume (e.g., $50k/year for 1M conversations). Volume discounts apply. No per-agent fees. No per-resolution charges. No add-on surcharges. What you see is what you pay. Multi-year agreements available. Dedicated CSM, premium support (4-hour response SLA), and custom onboarding included at Enterprise tier. No vendor lock-in: if you outgrow Rasa, export your models and dialogue definitions; they are plain YAML and JSON.

