Can Agentforce and ServiceNow AI agents work together?

Yes. Many enterprises use both platforms for their respective domains — Agentforce for customer-facing sales and service workflows, ServiceNow for internal IT and employee workflows. Integration between the platforms is possible through REST APIs or purpose-built connectors (ServiceNow has a Salesforce spoke in its Integration Hub). However, building and maintaining cross-platform agent workflows adds complexity; most organizations clearly define which platform owns which workflow domain rather than creating deeply interwoven agent chains across both.

How does Agentforce pricing compare to ServiceNow AI?

Salesforce Agentforce pricing is consumption-based: $2 per conversation for autonomous agent interactions (with volume discounts for committed usage). ServiceNow Now Assist is licensed per user and bundled with specific ServiceNow SKUs (ITSM Pro Plus, HR Service Delivery Pro, etc.) at an incremental cost above the base platform license. Organizations that already pay for ServiceNow enterprise licenses may find Now Assist more economical since it's included or incremental. Agentforce's consumption model can be cost-efficient for seasonal or variable-volume customer service workloads.

Which platform is easier to deploy AI agents on?

Both platforms require significant existing investment to get full value from their AI agent capabilities. Agentforce requires a configured Salesforce org with clean CRM data — agents work best when they have access to accurate customer records, case history, and product catalog data. ServiceNow Now Assist requires a properly configured ServiceNow instance with mature ITSM processes, a populated CMDB, and well-defined workflow templates. Neither is fast to deploy from scratch. Teams that have mature platforms get value quickly; teams building the underlying platform simultaneously will find AI deployment difficult.

Do either Agentforce or ServiceNow AI agents support custom LLMs?

Salesforce Agentforce is built on the Einstein platform with Salesforce's own models and select partner model options. It is not an open framework — you work within Salesforce's model choices. ServiceNow has announced model flexibility options, allowing enterprises to bring their own Azure OpenAI or other enterprise LLM connections in some configurations. Neither platform provides the model flexibility of open-source frameworks like LangChain or CrewAI. For enterprises with data sovereignty requirements or specific model preferences, these constraints matter and should be validated with each vendor before commitment.

Circuit board microchip representing enterprise AI processing power — Photo by JJ Ying on Unsplash

Salesforce Agentforce vs ServiceNow AI Agents (2026)

Enterprise AI agents in 2026 don't live in standalone tools — they live inside the platforms where work already happens. For customer-facing teams, that's often Salesforce. For IT and operations teams, that's often ServiceNow. Both companies have responded to this reality by building AI agent capabilities directly into their platforms.

Salesforce Agentforce and ServiceNow Now Assist are the respective results. Both enable AI-powered automation within existing enterprise workflows. Both target large organizations with established platform investments. And both require honest assessment of whether the AI capability is strong enough to justify building on the platform's constraints rather than choosing a more flexible alternative.

This comparison covers what each platform does well, where each falls short, and the organizational factors that should drive your choice.

For context on the broader enterprise AI landscape, see the Salesforce Agentforce profile and our comparisons index.

Quick Verdict#

Choose Salesforce Agentforce when your AI agent use cases are centered on customer relationships — sales assistance, case deflection, service resolution — and your team already lives in Salesforce.
Choose ServiceNow Now Assist when your AI agent use cases are centered on internal operations — IT incident resolution, HR case management, employee self-service — and your organization runs on ServiceNow.

Salesforce Agentforce Overview#

Salesforce Agentforce, announced at Dreamforce 2024 and GA in early 2025, is Salesforce's platform for building autonomous AI agents that take action within Salesforce and connected systems. It replaces and supersedes Salesforce's earlier Einstein Copilot offering with a more capable agentic architecture.

Agentforce agents are built on Topics (areas of expertise the agent handles), Actions (what the agent can do — Flows, Apex code, API calls, knowledge retrieval), and Instructions (natural language guidance for agent behavior). Agents operate within the Salesforce data model, meaning they natively access leads, contacts, cases, opportunities, products, and custom objects without integration work.

Key use cases:

Sales Development: Autonomous lead qualification, meeting scheduling, follow-up cadences
Customer Service: Case deflection, resolution recommendations, escalation routing
Commerce: Order status, returns, product recommendations
Field Service: Work order management, technician dispatch support
Marketing: Campaign performance analysis, audience segmentation recommendations

Agentforce's primary strength is depth of CRM integration — agents that need to read and write Salesforce data, trigger Flows, and update records have a seamless path with no integration layer required.

ServiceNow Now Assist Overview#

ServiceNow Now Assist is ServiceNow's generative AI capability, embedded throughout the ServiceNow platform across IT Service Management (ITSM), HR Service Delivery (HRSD), Customer Service Management (CSM), and other modules. Now Assist encompasses both copilot-style assistance (helping humans do work faster) and more autonomous agent behaviors.

ServiceNow's AI agent capabilities have evolved through 2025 with the introduction of AI Agents as a distinct offering — autonomous agents that handle entire workflows end-to-end rather than just augmenting individual tasks. These agents operate within ServiceNow's workflow engine, accessing the Configuration Management Database (CMDB), service catalog, and existing automation flows.

Key use cases:

IT Service Management: Incident classification, automated resolution for known issue patterns, change risk assessment
HR Operations: Employee onboarding workflow automation, policy Q&A, benefit inquiry resolution
IT Operations Management: Anomaly detection, incident correlation, automated remediation playbooks
Customer Service Management: Case classification, knowledge article recommendation, resolution automation
Procurement: Purchase request handling, vendor query resolution

ServiceNow's primary strength is deep process automation — AI agents that plug into mature ITSM and HRSD workflows, access rich operational data in the CMDB and service catalog, and take action through ServiceNow's existing automation framework.

Feature-by-Feature Comparison#

| Feature | Salesforce Agentforce | ServiceNow Now Assist / AI Agents | |---|---|---| | Primary domain | CRM, customer-facing | ITSM, HRSD, internal operations | | Agent building interface | Agent Builder (low-code) | AI Agent Studio (visual + code) | | Data model integration | Salesforce objects natively | ServiceNow CMDB + platform natively | | Process automation | Salesforce Flow + Apex | ServiceNow Flow Designer + scripts | | Knowledge integration | Salesforce Knowledge | ServiceNow Knowledge Management | | Pricing model | Per conversation ($2/conversation) | Per user, bundled with platform SKU | | External system integration | MuleSoft, REST, platform events | Integration Hub, REST, spoke library | | Governance / guardrails | Einstein Trust Layer | Now Assist policies + AI governance | | Multi-agent coordination | Agent Network (orchestrator pattern) | Multi-agent workflows (GA 2025) | | Customization depth | Topics, Actions, Flows, Apex | Flows, scripts, Spoke actions |

Pricing Comparison#

Salesforce Agentforce:

$2 per conversation (autonomous agent interaction)
Volume discounts available with committed usage agreements
Requires Einstein Platform license (included in many Salesforce clouds)
Salesforce CRM licenses remain a prerequisite

ServiceNow Now Assist:

Incremental cost added to base ServiceNow platform license
Bundled with specific Pro Plus or Enterprise SKUs (pricing negotiated with ServiceNow)
Example: ITSM Pro Plus adds Now Assist for IT use cases above ITSM Pro pricing
Not available as a standalone purchase — requires existing ServiceNow platform investment

Organizations with large, established ServiceNow contracts often find Now Assist economical because it's available at an incremental uplift to licenses they already own. Agentforce's per-conversation model rewards low-volume or highly variable workloads — you don't pay for capacity you don't use. At high customer service volume (millions of conversations), Agentforce's consumption model warrants careful cost modeling against fixed pricing alternatives.

Developer Experience#

Agentforce is built for Salesforce administrators and developers. Configuration happens in Agent Builder, a low-code interface within Salesforce Setup. Defining Topics and Actions requires understanding Salesforce's Flow and Apex ecosystem. Teams already experienced in Salesforce customization can build Agentforce agents without learning new tools. Teams new to Salesforce development face a steeper learning curve.

ServiceNow Now Assist / AI Agents configuration happens within ServiceNow's existing administration tooling plus the AI Agent Studio introduced in 2025. Teams familiar with ServiceNow Flow Designer, catalog items, and scripting can extend AI agents naturally. The integration with existing ITSM process configuration means AI agents often require coordination between AI configuration and ITSM process owners.

Both platforms require platform-specific knowledge that doesn't transfer between them. A Salesforce developer can't easily move to ServiceNow configuration, and vice versa. The developer experience evaluation is inseparable from existing team expertise.

When to Choose Salesforce Agentforce#

Agentforce is the right choice when:

Your AI agent use cases are customer-facing — sales qualification, service resolution, commerce support, field service
Your team already works primarily in Salesforce and benefits from zero-friction CRM data access
You want rapid deployment of agents that work directly on Salesforce records without integration work
Consumption-based pricing fits your workload profile — particularly for variable-volume customer interactions
You need strong governance controls through the Einstein Trust Layer, which ensures customer data doesn't leave Salesforce's infrastructure
Your existing automation is built on Salesforce Flow and you want AI agents to extend that investment

For perspective on how platform-native AI agents compare to more flexible frameworks, see the open-source vs commercial AI agent frameworks guide.

When to Choose ServiceNow Now Assist#

ServiceNow Now Assist fits best when:

Your AI agent use cases are internal operations — IT incident resolution, employee self-service, HR workflows, procurement
Your organization has a mature ServiceNow deployment with a populated CMDB, established ITSM processes, and a service catalog
You've negotiated enterprise ServiceNow licensing that makes Now Assist an incremental uplift rather than net-new spend
You need AI agents to work within existing ITSM workflows — incident lifecycle, change management, problem management
Your employees interact primarily through the ServiceNow employee portal rather than external customer channels
You need AI-assisted IT operations management — correlating alerts, suggesting remediation, automating known resolutions

Verdict#

The comparison between Agentforce and ServiceNow Now Assist is largely a comparison between customer-facing and internal-facing enterprise AI automation — two different problems that happen to use similar underlying technology.

Most large enterprises will use both: Agentforce to automate customer interactions that flow through CRM, and ServiceNow to automate internal IT and HR operations that flow through ITSM. The platforms don't compete for the same workflows in the same way that, say, CrewAI and LangChain compete for the same developer mindshare.

The meaningful questions for enterprise decision-makers are: Which platform is more mature in your organization? Where is your data better organized? Which team owns the implementation? And critically — is the platform's AI capability strong enough to justify the constraints of a closed, proprietary system versus building on a more flexible agent framework?

For teams with mature Salesforce and ServiceNow deployments and IT and sales operations teams familiar with each platform's administration model, the answer is often yes — the integration benefit outweighs the flexibility cost. For teams building net-new, the open-source alternatives deserve serious evaluation.

Frequently Asked Questions#

Can Agentforce agents access data outside of Salesforce?

Yes. Agentforce agents can make external API calls via External Services or direct REST calls configured as Actions. MuleSoft integration enables Agentforce agents to access data from ERP systems, databases, and other enterprise applications. However, each external data source requires explicit integration work — the zero-friction data access applies to Salesforce objects only. For use cases requiring deep integration with non-Salesforce systems, evaluate whether the integration overhead is worth the Agentforce benefits.

How does ServiceNow Now Assist handle sensitive HR or IT security data?

ServiceNow Now Assist operates within ServiceNow's existing security and access control framework — agents can only access data that the configured service account has permission to view. Sensitive HR data (compensation, performance) and security data (vulnerability details, audit logs) are governed by ServiceNow's ACL (Access Control List) system. The Einstein-equivalent governance model in ServiceNow ensures that AI-generated responses respect data access boundaries. For highly sensitive data classes, ServiceNow recommends explicit exclusion rules in Now Assist configuration.

What happens to Agentforce or ServiceNow AI agents if the underlying LLM changes?

Both platforms abstract the LLM from the agent configuration — you don't write prompts that depend on specific model characteristics. Salesforce manages model updates within the Einstein platform; ServiceNow manages model updates within Now Assist. In practice, this means vendors can improve or swap the underlying model without breaking configured agents (though behavior may change subtly). The abstraction is a feature of platform-native AI — you don't manage model versions, but you also don't control them.