“潮流前线 (What's Trending)”的CEO和创始人切丽尔·拉扎尔 (Cheryl Lazar) 在“AI下载 (The AI Download)”节目中采访了 Slack 的首席营销官 (CMO) 瑞安·加文 (Ryan Gavin)。讨论的重点围绕 Slack 新推出的 AI 驱动的协作功能，AI 如何重塑劳动力市场，以及公司如何适应这一快速变化的格局。 这一集首先快速浏览了 AI 头条新闻，包括： * 一个 AI 产生幻觉的案例，律师提交的法院文件中包含 AI 生成的虚假法律引文，突显了验证 AI 生成信息的重要性。 * 欧盟的 AI 行为准则，重点关注版权和安全，特别是针对使用公共互联网数据训练大型语言模型 (LLM) 的公司。加文澄清说，Slack 的方法不同，因为他们专注于客户数据，这些数据仍然是客户的财产，确保数据的安全性、安全性和隐私性。Slack 绝不会使用您的数据来训练 LLM。 * Indeed 和 Glassdoor 因为 AI 对求职业务的影响而裁员。加文认为 AI 不会接管企业或工作，但工作将会发生变化，他认为这种变化主要在于提升人类的能力，特别是对于小型企业而言，它们现在可以利用 AI 来像大型企业一样运作。 * Reddit 帖子重新点燃了关于 AI 在医疗建议中作用的争论，引用了一个 ChatGPT 标记了一个医生十多年来都未发现的基因突变的案例。 然后，主要焦点转移到 Slack 的新 AI 集成，这些集成由 Salesforce 的 Einstein AI 平台提供支持。加文强调了 AI 代理在 Slack 中的集成，利用该平台庞大的非结构化数据（对话、消息、文件）存储库，为代理提供更深入的背景信息和见解。这些代理可以利用公司的长期记忆，访问历史数据来改进他们的决策和建议。 加文强调了隐私和权限的重要性，向听众保证 Slack 中的代理继承用户权限，确保他们只能访问用户有权查看的数据。交互设计最初是私有的，允许用户在频道中公开分享代理的响应之前对其进行审查。 EasyCator，一家餐饮公司，被用作企业如何使用它的一个例子。他们在 Slack 中部署了一个代理来处理传入的客户请求。另一家公司 Anthropic，使用 Salesforce 和 Slack，交易处理速度提高了 60%。 在谈到易用性时，加文解释说，Salesforce 的代理力量平台是低代码的。公司管理员可以将代理部署到 Slack，以用户友好的方式提供给用户，就像一个普通的队友一样。他引用了“别让我思考”的原则，强调 AI 与工作流程的无缝集成。 加文随后讨论了新数据，显示 Slack 中 AI 的每日使用量激增 233%，用户满意度提高了 81%。他认为，AI 可以将员工从“忙碌的工作”转变为“深入的工作”，从而应对因分散的工作流程和信息过载而导致的生产力停滞。他强调了在 AI 时代重新构想流程的重要性，并鼓励公司创建安全的空间来进行实验和创新。 在谈到在拥挤的市场中营销 AI 的挑战时，加文强调了 Slack 的方法，即将原生 AI 功能直接集成到平台中，通过不同的定价计划使其可访问。他重点介绍了 AI 驱动的企业搜索和 AI 频道摘要等功能，强调了这些集成的无缝、“隐形”性质。 展望未来，加文预测劳动力将从“执行者”转变为“协调者”。他认为 AI 将使个人能够协调工作，利用代理来自动化任务、生成内容和简化流程。他建议那些担心这种变化的人“保持好奇心”，尝试 AI 工具，并积极探索它们的潜在应用。 加文最后倡导一种好奇和积极的心态，强调那些拥抱实验并寻求新可能性的人将在不断发展的工作世界中蓬勃发展。

Cheryl Lazar, CEO and founder of What's Trending, hosts Ryan Gavin, the CMO of Slack, on "The AI Download." The discussion centers around Slack's new AI-powered collaboration features, how AI is reshaping the workforce, and how companies can adapt to this rapidly changing landscape. The episode begins with a quick look at AI headlines, including: * A case of AI hallucination where lawyers submitted court filings with fake legal citations generated by AI, highlighting the importance of verifying AI-generated information. * The EU's AI code of practice, focusing on copyright and safety, especially for companies training LLMs on public internet data. Gavin clarifies that Slack's approach differs, as they focus on customer data that remains the customer's property, ensuring data safety, security, and privacy. Slack will never train an LLM on your data. * Layoffs at Indeed and Glassdoor due to AI impacting the job search business. Gavin doesn't think AI is taking over businesses or jobs, but that work is going to change, and he believes this change is primarily about upgrading human capabilities, especially for small businesses which can now leverage AI to act like larger corporations. * Reddit posts reigniting debates about AI's role in medical advice, citing an instance where ChatGPT flagged a genetic mutation missed by doctors for over a decade. The main focus then shifts to Slack's new AI integrations, powered by Salesforce's Einstein AI platform. Gavin highlights the integration of AI agents within Slack, leveraging the platform's vast repository of unstructured data (conversations, messages, files) to provide agents with deeper context and insight. These agents can tap into a company's long-term memory, accessing historical data to improve their decision-making and recommendations. Gavin emphasizes the importance of privacy and permissioning, assuring listeners that agents within Slack inherit user permissions, ensuring that they only access data the user is authorized to view. The interactions are designed to be private initially, allowing users to review the agent's response before sharing it publicly within a channel. EasyCator, a business catering company, is given as an example of how businesses can use this. They deployed an agent in Slack to handle inbound customer requests. Another company, Anthropic, uses Salesforce and Slack and has seen a 60% improvement in the speed of processing deals. Addressing the ease of use, Gavin explains that Salesforce's agent force platform is low-code. Company admins can deploy the agents to Slack, making them available to users in a user-friendly manner, just like a regular teammate. He cites the principle of "don't make me think," emphasizing the seamless integration of AI into the workflow. Gavin then discusses new data showcasing a 233% surge in daily AI use within Slack and an 81% increase in user satisfaction. He argues that AI can shift workers from "busy work" to "deep work," combating the productivity stagnation caused by fragmented workflows and information overload. He stresses the importance of reimagining processes in the age of AI and encourages companies to create safe spaces for experimentation and innovation. Addressing the challenge of marketing AI in a crowded marketplace, Gavin emphasizes Slack's approach of integrating native AI capabilities directly into the platform, making them accessible through different pricing plans. He highlights features like AI-powered enterprise search and AI recaps of channels, emphasizing the seamless, "invisible" nature of these integrations. Looking ahead, Gavin predicts a shift in the workforce from "doers" to "orchestrators." He believes AI will empower individuals to orchestrate work, leveraging agents to automate tasks, generate content, and streamline processes. He advises those concerned about the changing landscape to "be curious," experiment with AI tools, and proactively explore their potential applications. Gavin concludes by advocating for a curious and proactive mindset, emphasizing that those who embrace experimentation and seek new possibilities will thrive in the evolving world of work.