
销售的未来是代理驱动的。我们谈论的是 AI 代理自动化外联、分析 CRM 数据以获取洞察,甚至起草个性化提案。但关键在于:能力越大,责任越大——随之而来的是重大的安全挑战。CRM 领域无可争议的重量级冠军 Salesforce 正在正面应对这一问题,销售领导者需要密切关注。
他们最近的洞察强调了一个关键事实:“前沿模型加上代理式框架会暴露出比工程团队所能吸收的更多真实、可利用的漏洞。” 对于销售部门来说,这意味着:您的 AI 代理越强大、越自主,它们在您的系统中可能暴露的潜在漏洞就越多。这不仅仅是一个后端 IT 问题;它直接威胁到您的销售渠道、客户关系,并最终影响您的季度配额。
试想一下。您的 CRM 是销售运营的心脏,是敏感客户数据的金矿。当您授权 AI 代理与这些数据交互时,风险会急剧增加。一次安全漏洞不仅仅是修复一个 bug;它可能是一次数据泄露,会侵蚀客户信任,引发昂贵的合规罚款,并扼杀您的交易流程。如果安全措施不牢固,您从 AI 自动化中追求的投资回报率可能会比未达成的配额消失得更快。
好消息是?像 Salesforce 这样的行业巨头不仅仅是在观察问题;他们正在积极构建防护措施。他们专注于“大规模管理代理安全”,这意味着他们正在创新,以确保您集成到销售技术栈中的 AI 工具——尤其是那些深度嵌入 CRM 的工具——能够抵御新兴威胁。这种积极主动的姿态对于希望在不损害诚信的情况下利用 AI 的销售团队来说是一个巨大的胜利。
对于销售领导者而言,这意味着一个明确的任务:要求您的 AI 供应商提供强大、针对代理的特定安全框架。不要让安全担忧阻碍您的自动化雄心。相反,选择那些优先考虑这些高级安全措施的平台,使您的团队能够自信地部署 AI 代理。当您的 AI 工具安全时,您的销售代表可以信任数据,加速交易周期,并专注于他们最擅长的事情:成交。这不仅仅是修补代码;这是关于保护和推动您的收入流。那些拥抱安全、智能自动化的团队将是那些持续达到——并超越——其目标的团队。
图片:DJI-Agras / Pixabay (https://pixabay.com/photos/agriculture-drone-dji-agriculture-4208863/)
Meta launches Muse Wonder while Jev raises $870M, reshaping AI‑driven sales automation and the broader AI ecosystem.

Forget spray-and-pray event strategies. Smart sales leaders are zeroing in on AI-focused gatherings to generate massive leads and fuel their 2027 pipelines, demanding real ROI.

SaaStr's new AI Agent Playbook reveals how they run Go-To-Market with just 3 humans and 21+ AI agents, offering battle-tested strategies for sales leaders to boost efficiency and hit quotas.

AI agents are now evaluating tech stacks and picking vendors autonomously. Sales leaders must adapt their pipeline strategies for autonomous B2B buyers.

评论 (2)
How do you think Salesforce's approach to agentic security will impact the adoption of AI agents in smaller sales teams with limited IT resources?
Salesforce’s baked‑in security controls actually level the playing field—small teams can deploy agents without hiring a full‑time security admin, and the out‑of‑the‑box compliance checks shave weeks off onboarding, driving faster time‑to‑revenue. The trade‑off is that you’ll need to lean on Salesforce’s managed‑service add‑ons or a trusted partner to keep the cost of oversight under control, otherwise the ROI can evaporate.
Great point on the security‑risk trade‑off; when an AI agent unintentionally exposes CRM data, the fallout isn’t just a technical incident—it can instantly drop CSAT scores and drive up support tickets, eroding the very deflection gains we strive for. Have you seen any early metrics on how sales‑focused breaches impact post‑sale NPS or ticket volume, and what mitigation steps can preserve both pipeline velocity and customer trust?
I’ve seen early data from a handful of firms where a sales‑focused breach shaved roughly 12 % off post‑sale NPS and spiked support tickets by about 30 % within two days. The fix is to lock agents to the minimum field‑level permissions, enable real‑time anomaly alerts, and run just‑in‑time access reviews so you keep pipeline speed while protecting trust.