In a recent interview with Dong-A Ilbo, Park Se-jin, CEO of Salesforce Korea, defined Salesforce's AI agent platform 'Agentforce' as 'digital labor' and stated this. Park emphasized, "AI will never leave our side, so we must find ways to coexist," adding, "AI should replace simple repetitive tasks, and humans should be assigned to more strategic and creative tasks."

This aligns with the recent declaration by Marc Benioff, CEO of Salesforce, on a podcast, stating, "We are no longer just a software company providing information management tools. We are now offering digital workers." Recently, Elon Musk, CEO of Tesla, also wrote on his social media, "AI and robots will replace all jobs. Labor will become optional."

The case of Salesforce alone shows that the labor market transformation due to AI is already underway. Benioff revealed, "Salesforce's customer support department reduced its workforce from 9,000 to about 5,000," and stated, "Currently, AI performs 30-50% of all tasks at Salesforce." He declared, "We must now accept the reality that AI will take over tasks previously done by humans, and we will move to higher value-added work."


●‘Self-working AI Employee’… Unlike existing chatbots, solves problems independently without human intervention

Agentforce is Salesforce's AI agent platform. Simply put, it can be described as a 'self-working AI employee.' Unlike existing chatbots that follow pre-defined question-answer patterns, Agentforce can flexibly respond according to the situation, understand conversations like a human, plan, and act independently.

For example, if an existing chatbot only provides prepared answers to the inquiry 'I can't log in,' Agentforce can directly check the customer's account status, identify the cause of the problem, and resolve it by directly connecting to the relevant department's practitioner if necessary. The feature is that this process does not require human intervention. It can handle multiple requests simultaneously around the clock and accurately manage structured data. For instance, in hotel reservations, while a human might need 10-20 exchanges, the agent can understand and process the customer's intent and conditions in one go.

Within a year of its launch, Agentforce was adopted by 8,000 companies worldwide, with half of them, 4,000 companies, using it beyond proof of concept (PoC) to paid use. Domestically, Cafe24, Line Pay Plus, and Hyundai Infracore have begun actual application, and currently, about 100 companies are considering adoption.

According to a study conducted by Salesforce and global market research firm IDC with over 2,000 experts in North America, Europe, and Asia, AI-based 'Digital Labor' is expected to generate approximately USD 13 trillion (about KRW 18,600 trillion) in economic effects by 2030. The study analyzed that AI agents could support human decision-making and automate 22% of employee tasks to enhance productivity.


● The era of ‘Agent to Agent (A2A)’ collaboration begins in earnest

Park stated, "Many inquiries from clients ask whether digital labor can expand customer support handled by 100 employees to seem like 1,000 employees working at once."

Formula One (F1) is cited as a representative case that adopted Agentforce, reducing average response time by 80% and call handling time by 50%. After the global success of the racing movie 'F1 The Movie' starring Brad Pitt, the fan base rapidly increased, making fan management innovation crucial for F1. This was because a new fan base emerged rapidly, including not only existing fans but also the general public previously uninterested in the races.

Among F1's entire fan base, those under 35 account for one-third, and the proportion of female fans increased from 32% in 2018 to 42% in 2023. Social media followers exceeded 96 million. Less than 1% of fans watch the Grand Prix live, but most fans engage with the races through digital and indirect experiences such as F1 TV, games, and merchandise purchases. In this process, over 100 internal and external data points are generated. F1 integrates and analyzes this large-scale data to design personalized experiences for fans and increase loyalty.

The first consultation problem resolution rate for Agentforce for F1 is over 95%, and the click-through rate for AI-based content recommendations increased by 22%. F1 integrated 100 data sources to build profiles for 24 million fans and provided customized services. The annual fan satisfaction rate reached 90%. F1 plans to expand its fan base to 43 million by 2027. An F1 official stated, "Through Agentforce, we are strengthening connections with fans and enhancing fan experiences at all touchpoints."

Salesforce is also pursuing the construction of an 'agent ecosystem' where multiple agents collaborate beyond a single agent. In this ecosystem, multiple agents with different goals cooperate to perform complex tasks.

For example, in car transactions, the seller's agent aims for the highest price, while the buyer's agent aims for the lowest price, but they recognize and coordinate each other's goals to conclude the transaction. Performing without errors in such 'Mixed Objective Scenarios' is a key challenge for A2A technology. If applied to distribution, the retailer's agent receives a consumer's return request, the manufacturer's agent assesses product damage, and the logistics partner's agent determines the delivery procedure.

● Global AI Readiness Index assessment… Korea's AI readiness ranks among the world's top

According to the 'Global AI Readiness Index' released by Salesforce in July 2024, Korea scored 9 out of 10 in the AI regulatory framework category, ranking among the top alongside Singapore and the UK. The national AI strategy and data governance system received high evaluations.
The level of AI adoption by industry and government institutions scored 6.7, with rapid AI application expansion in key industries such as manufacturing, smart cities, and logistics being noted as strengths.

Park explained, "Korea has relatively strong regulations, but human capital and technology levels are highly rated compared to global standards," adding, "Most indicators exceeded the global average, and considering the size of the national economy, the AI readiness index is very high."

He also emphasized, "Korea is quick to adopt AI and responds swiftly," adding, "The headquarters is expanding investments as it is a market gaining attention due to manufacturing infrastructure like semiconductors and the spread of K-culture. The headquarters is particularly positively interested in the Korean government's policy to promote sovereign AI."

Popular News