Understanding the transformative power of agentic AI

The figures speak volumes: Grand View Research predicts the global AI agents market will surge from $5 billion in 2024 to $50 billion by 2030, representing a 46% annual growth rate. Even more significantly, research from the Capgemini Research Institute suggests that AI agents could create up to $450 billion in economic value by 2028 through cost reductions and revenue generation. However, despite this vast potential, only 2% of organizations have fully implemented AI agents, leaving a limited timeframe for businesses to gain a competitive edge before the market becomes saturated.

Unlike conventional AI systems that simply respond to prompts, agentic AI exhibits true “agency” — the ability to define objectives, make decisions, and act independently with little human intervention. Harvard Business Review characterizes these systems as possessing "enhanced reasoning and execution abilities" that go well beyond answering questions to actually completing complex operations.

This distinction is vital: while generative AI focuses on generating content through language-to-language interactions, agentic AI is centered on multi-step reasoning, planning, and most importantly, action. It can arrange travel, process insurance claims, manage stock levels, and perform in-depth research across hundreds of sources. This level of autonomy shifts AI from being a tool to becoming a genuine digital partner.

What Exactly Constitutes an AI Agent?

Unlike conventional AI that reacts to prompts, an AI agent is an artificial intelligence system capable of managing multi-step tasks without continuous human involvement. This marks the next evolution in AI — “Agentic AI.” While ChatGPT provides answers, AI agents actually perform actions — booking flights, processing invoices, debugging code, and conducting independent research across numerous sources.

The defining trait: agents can execute multiple steps, interface with various applications, and operate over extended durations. For example, OpenAI's Codex agent can function for up to 30 minutes without human input, and Anthropic's Claude 4 can handle coding tasks continuously for seven hours.

Seven Categories of Digital Workers

Although there will eventually be countless AI agents, we can currently classify them into distinct types that are beginning to enter the workforce. The Information offers a helpful breakdown of the different forms of digital labor:

1. Business-Task Agents

What they do: Manage enterprise processes across multiple software platforms Digital labor: Invoice handling, data input, document sorting, scheduling Examples: UiPath, Microsoft Power Automate, Zapier AI

2. Conversational Agents

What they do: Address customer and employee inquiries via dialogue Digital labor: Customer support, IT helpdesk, HR functions
Examples: Salesforce Agentforce, ServiceNow NowAssist, Sierra, Decagon

3. Research Agents

What they do: Gather, evaluate, and verify data from credible sources Digital labor: Academic studies, citation verification, technical evaluations
Examples: OpenAI Deep Research, Perplexity Pro, Scite Assistant, AlphaSense

4. Analytics Agents

What they do: Interpret data to generate visual reports and summaries
Digital labor: Data querying, dashboard creation, business intelligence
Examples: Power BI Copilot, Tellius, ThoughtSpot, Glean

5. Developer Agents

What they do: Assist software engineers with complex coding tasks
Digital labor: Code suggestions, debugging, documentation, system maintenance Examples: Cursor, GitHub Copilot, Claude Code, Cognition's Devin

6. Domain-Specific Agents

What they do: Deliver specialized performance in regulated fields like law, healthcare, and finance
Digital labor: Legal document review, medical triage, financial assessments
Examples: Harvey (legal), Hippocratic AI (healthcare), Rogo and Hebbia (finance)

7. Browser-Interacting Agents

What they do: Navigate websites and perform repetitive online tasks
Digital labor: Online form submissions, web ordering, social media management
Examples: OpenAI Operator, Google Project Mariner, Anthropic Computer Use

OpenAI’s ambitious vision comes to life

OpenAI’s rollout of AI agents started in January 2025 with Operator, an AI capable of interacting with websites like a human — clicking buttons, completing forms, and navigating pages. This was followed in February by Deep Research, which synthesizes insights from hundreds of sources to produce fully referenced reports within minutes. By July, the release of ChatGPT Agent consolidated these capabilities into what The Wall Street Journal describes as "an agent that can create spreadsheets and presentations" while managing complex, multi-step workflows.

Sam Altman, CEO of OpenAI, anticipates these agents will "significantly impact company performance" in 2025, estimating they can already perform "a low single-digit percentage of all economically valuable work globally." With a 41.6% accuracy rate on complex reasoning benchmarks — twice that of earlier models — these agents mark a major advancement in AI capability.

Redefining consumer and business experiences

AI agents are transforming both consumer engagement and business operations on an unprecedented scale. For consumers, the shift is happening rapidly: recent reports indicate that a growing number of customer interactions in 2025 are being managed by AI, with current implementations showing AI-powered systems reducing resolution times and improving customer satisfaction compared to traditional support.

The impact on consumers goes beyond convenience. Klarna’s AI assistant cut average issue resolution from 11 minutes to just 2 minutes, maintaining customer satisfaction levels comparable to human agents. Virgin Money’s AI assistant "Redi" has managed over 2 million interactions with a 94% satisfaction rate, proving that consumers are ready to embrace AI-powered service when it delivers superior results. Retail adoption is equally impressive, with 24% of shoppers already comfortable allowing AI agents to make purchases — a figure rising to 32% among Gen Z, while 75% of customer inquiries can now be resolved by AI without human involvement.

The business benefits of AI agents are equally strong, supported by real-world results. Organizations using AI report an average revenue increase of 6-10%, with 62% of companies expecting full or even greater ROI. Operational improvements are striking: 83% of companies using AI report revenue growth compared to 66% without AI, 76% see improved operational efficiency, and financial institutions benefit from increased profitability through better fraud detection and personalized services.

Real-world success stories highlight the transformation across industries. JPMorgan Chase’s AI-powered "Coach" tool helps financial advisors access research 95% faster, contributing to a 20% increase in asset management sales. The bank’s AI initiatives have already saved nearly $1.5 billion through fraud prevention and operational improvements. Wiley achieved a 40% increase in case resolution using AI agents, while 76% of e-commerce teams credit AI with boosting sales and 92% of service teams report cost reductions. Manufacturing leaders have seen a 40% decrease in downtime thanks to AI-powered predictive maintenance.

Improvements in employee productivity are equally notable, with customer service agents handling more inquiries per hour, business professionals producing more documents per hour, and developers completing more coding projects weekly with the help of AI agents. These are just early applications, but they clearly show how agentic AI will reshape the standards of customer experience and business performance.

Why This Is Just the Start

We're in the early stages of the digital labor revolution. Current agents still make errors and require oversight, but their capabilities are advancing rapidly. With more affordable reasoning models, improved orchestration tools, and expanding integrations, the capabilities of AI agents are growing exponentially.

The workforce of 2030 will not consist solely of humans — it will be a hybrid environment where digital agents manage repetitive tasks while humans focus on creativity, strategy, and relationship-building. We're not merely automating work; we're introducing a new class of digital colleague that enhances human ability rather than replacing it.

The era of digital labor has arrived. The question is no longer whether AI agents will change the workplace — it's how quickly businesses and consumers will adapt to this new reality.

The above is the detailed content of From Tools To Teammates: How AI Agents Will Become Digital Labor. For more information, please follow other related articles on the PHP Chinese website!