推动盈利增长的AI创新方法：生成式AI与AI智能体变现指南

A how-to guide to monetizingGenAI and AI agents Contents Welcome to the Agentic AI era How to drive profitable revenue growth through AI innovation8 Outcome-based pricing & packaging How to adjust operations for agentic AI and GenAI33 Turning into action Welcome to the Agentic AI era A year ago, most conversations were about large language models, foundation models,or copilots. These were still framed as assistive tools: things you prompt, things thataugment human effort, things that sit alongside you. Now the narrative is shifting: notjust tools that help, but agents that act. Instead of being reactive to prompts, agents can The projections are staggering, almost doubling the size of the market in just four or fiveyears. Agents promise to swallow entire categories of repetitive, operational work, This shift to agents isn’t happening in a vacuum. It’s the next step in a much longerjourney of how businesses adopt and value software. Each prior era reshaped not only the In the on-premise era, software was essentially infrastructure. Customers bought bigsystems, paid heavy upfront CAPEX, and installed them in their own servers. They wereback-office enablers, not frontline drivers of growth. Scalability was limited, innovation Then came the SaaS era, roughly from 2010 through today. This was the breakthrough thatturned software into a utility customers could scale flexibly. Instead of massive upfrontinvestment, they subscribed. Instead of running their own servers, they tapped the cloud.Software shifted from back-office to the center of business operations: CRM, ERP, HR, The software market saw strong growth when moving from Perpetuallicences to ‘software-as-as-service’; imagine how much it couldexpand shifting to ‘Services-as-a-software’, capturing a portion of the business services market on top of its current TAM Now we’re on the cusp of the AI Agent era. What’s different here is that software is nolonger just a tool. Agents can autonomously execute workflows, make recommendations,aand, if they are sophisticated, take actions on their own. The growth projections showhow transformative this is expected to be: the market could almost double in just four or This new era is also redefining value. In the on-premise era, value was tied to infrastruc-ture. In the SaaS era, value was tied to access and scalability. In the AI Agent era, value You may expect this paper to tell a tale of technology. However, we reframe this develop-ment as a business model story. Because as agents really do start mirroring or augmen-ting human workflows, then the way companies buy and pay for software will have to Why does the AI agent era feel so different? There’s a reason why the market projections are so aggressive. Up to now, automationwas mostly about workflows (rules, triggers, moving information from one place toanother, etc.) However, cognitive work, the kind of thinking humans have traditionally That’s what agents are beginning to crack. They can support decision-making in areassuch as sales outreach and customer engagement, tackle complex problem-solving The second driver is personalization at scale. Humans can only handle so much tailoringbefore it becomes overwhelming. AI agents, on the other hand, can customize interac-tions for thousands of customers or employees simultaneously, in real time. Every email,every product recommendation, every HR policy reminder, all personalized. This changes The third factor is the data explosion. Enterprises are drowning in data they can’t fullyuse. Legacy analytics tools help, but they’re slow, require expertise, and often leave insights buried. AI agents can chew through that data, find patterns, and surface actionsin seconds. That makes the value proposition hard to ignore, because companies know Finally, there are new revenue streams. AI-native applications bring fresh monetizationpaths: usage-based models tied to consumption, commission-style models tied toperformance, and embedded intelligence layered into existing SaaS platforms as Addressing the payment disconnect Modern employment is structured around renting time and skills through wages. Compensation has traditionally been tied either totime(hourly pay),performance(commissions), orscarce expertise(high hourly rates for specialists). The idea of compensating workers by the hour emerged during the Industrial Revolutionin the late 18th and early 19th centuries. Beforehand, many workers were paid by thepiece (piecework in agriculture or textiles) or through fixed contracts. As factoriesspread, employers needed a way to control and measure labor more consistently. From an employer’s mindset, this created a simple equation: more hours worked = morepay owed. Productivity could be adjusted by scheduling more or fewer hours. It’s a That same question now confronts the world of AI. As AI systems begin to take onhuman-like tasks, emulating or augmenting them, the old SaaS levers of “features and Why do we th