Archive · 16 Sep 2026
Anthropic Expands Claude-Salesforce Partnership Amid CEO's AI Safety Warnings
Anthropic is putting its Claude AI into Salesforce’s business software.
The expanded partnership is called Claudeforce.
It also connects Claude with Agentforce and Slack.
Workers can ask Claude questions about customers, sales deals and business forecasts.
Salesforce provides the company information and rules, while Claude helps reason through tasks.
At the same time, Anthropic CEO Dario Amodei has warned that powerful AI could become difficult for people to control.
He says companies should slow down frontier-AI development to allow more time for safety testing.
Amodei also says AI agents should work together with software companies rather than destroy them.
Anthropic expanded its Salesforce partnership to integrate Claude into Salesforce, Agentforce and Slack.
Salesforce in Claude is available in beta with 37 pre-built sales skills.
Users can ask Claude to analyze pipelines, prepare calls, create dashboards and generate forecasts.
Marc Benioff described Claude as probabilistic AI working with Salesforce’s deterministic systems and business data.
Dario Amodei has urged slower frontier-AI development while saying AI agents should work with, not replace, SaaS.
- Who
- Anthropic, led by CEO Dario Amodei, and Salesforce, led by CEO Marc Benioff.
- What
- The companies expanded their partnership to integrate Claude into Salesforce, Agentforce and Slack.
- When
- The expansion was reported shortly after Amodei published warnings about the risks of advanced AI; Salesforce in Claude is now available in beta.
- Where
- Within Salesforce’s enterprise software, Agentforce and Slack.
- Why
- The partnership is intended to let businesses use Claude’s reasoning capabilities with Salesforce data, workflows and business context, while Amodei argues that frontier-AI development should proceed more slowly for safety.
Enterprise AI Expansion
AI Safety and Disruption Concerns
How quickly AI should advance
Enterprise AI Expansion
The partnership expands the practical use of Claude in business software and makes advanced AI available for sales and workplace tasks.
AI Safety and Disruption Concerns
Dario Amodei says AI capabilities are advancing drastically faster and companies should slow frontier development to allow more time for alignment, security and testing.
AI’s relationship with SaaS
Enterprise AI Expansion
Salesforce and Anthropic present Claude as a reasoning layer that works with Salesforce data, workflows and business rules.
AI Safety and Disruption Concerns
The rise of AI agents has led some observers to describe them as potential SaaS killers, although Amodei rejects the idea that Anthropic wants to destroy SaaS companies.
Control of increasingly capable systems
Enterprise AI Expansion
Anthropic’s proposal supports continued AI development at what Amodei calls a balanced rate, alongside safeguards and evaluation.
AI Safety and Disruption Concerns
Amodei warns that future systems could help build the next generation of AI through recursive self-improvement, potentially moving faster than humanity can understand or control.
- Partnership name
- Claudeforce
- AI model
- Anthropic’s Claude
- Integrated products
- Salesforce, Agentforce and Slack
- Product status
- Salesforce in Claude is available in beta
- Pre-built capabilities
- 37 sales skills
- Slack integration
- Claude is becoming the default model for Slack AI and Slackbot
- Safety proposal
- Amodei has called for a balanced pace of frontier-AI development, including third-party evaluators and greater coordination