"AI Application" Changes: Copilot is Dead, Agent Rules?

Recently, Salesforce CEO Marc Benioff openly criticized Microsoft's Copilot AI at the company's annual Dreamforce conference, likening it to the failed digital assistant Clippy from the last century.

"We all know now that Microsoft Copilot is basically the new Microsoft Clippy, and customers are not getting value from it," Benioff told Bloomberg in an interview.

Clippy, launched by Microsoft in the 1990s, was a cartoon paperclip assistant that provided suggestions to users. Due to its high interference and poor utility, it was heavily criticized and eventually phased out by Microsoft in the early 2000s.

Benioff pointed out that many customers complain that Copilot is not as good as OpenAI's ChatGPT, even though both use the same technological foundation.

Tech Giants Betting on Agents

In stark contrast, Benioff expressed full confidence in Salesforce's latest AI agent product - Agentforce.

Agentforce is a type of AI agent designed to provide direct productivity enhancements for businesses and their employees without the need for companies to develop and adjust AI models themselves.

"We have attracted tens of thousands of customers and told them that they don't need to develop AI solutions themselves," Benioff said in the interview.

In fact, many global tech giants such as Microsoft, Workday, and other companies are pushing Agent systems to the core of their development strategies, aiming to not only provide assistance but also to complete tasks on behalf of users.

Microsoft CEO Satya Nadella stated that the ultimate form of Agent is an "enterprise orchestration layer" where employees can generate and use these agents through conversations with the system to complete various complex tasks.

Companies like Salesforce are targeting areas such as automated customer service, aiming to use AI agents to handle more complex work and even replace some positions.

Can AI Agents Truly Land?

Despite the optimism surrounding the future of Agents, this technology still faces many challenges.

One prominent issue is that generative AI systems are prone to "hallucinations," generating inaccurate information in uncertain situations, leading users to be cautious about relying entirely on Agents to perform tasks.

Former Microsoft CTO Barry Briggs expressed caution and nervousness about the widespread application of AI agents, believing that such systems must operate within workflows where employees have final decision-making authority.

Furthermore, software companies traditionally charged customers based on the number of employees using the software. However, with Agents replacing some human labor, this pricing model will also come under pressure