1. AI agents: “digital assistants that get the job done on their own once you assign it”
In a nutshell, an AI agent is software that, once you give it a goal, makes its own decisions and carries out the work for you.
It’s not a chatbot that just answers questions. It looks at its surroundings (data and systems), fetches the information it needs, decides on its own what to do next, and pushes the task through to the end.
For example, if you say, “Analyze my card spending and draw up next month’s budget,” it reads your account and card data, analyzes spending patterns by category, compares them with this month, and even proposes a sensible budget.
2. SaaS: “internet software you subscribe to monthly, with no installation”
SaaS (Software as a Service) is, in plain terms, a subscription software service you rent over the internet.
Instead of installing and maintaining programs on your own computer as in the past, the provider hosts the software on cloud servers and you access it through a web browser or app.
- Users: sign up and just pay a monthly or yearly fee
- Service providers: handle server operations, security and feature updates themselves
- Advantages: no installation, access from anywhere, and when your team grows you simply add accounts
Much of what we use today, from email services and collaboration tools to accounting software and online store solutions, is already SaaS.
3. Why are AI agents and SaaS getting attention at the same time right now?
1) “Too much work, too few people”
Companies want to cut repetitive work, and labor-cost pressure is mounting. AI agents can automate much of rule-based knowledge work such as data collection, report writing, email replies and customer support.
SaaS delivers these AI features as ready-to-use “packaged products,” so companies can adopt them without hiring lots of developers.
2) “After ChatGPT, the focus shifts from ‘talking’ to ‘doing’”
If 2023–2024 was about AI that answers well, like LLMs (ChatGPT and others), from 2025 the focus is shifting from “AI that talks well” to “AI agents that do the work for you”.
AI agents don’t just answer; they are “action-oriented AI” that log into internal systems, change data, place orders and even adjust schedules.
3) “Cloud + subscription models have slashed the cost of experimenting”
In the past, this kind of automation meant buying a solution for hundreds of millions of won, installing it on-premises (on your own servers) and customizing it.
Now a team can try SaaS with built-in AI features for a few tens of thousands of won a month and, if it works, roll it out company-wide right away.
The AI agent market is growing fast, too, expected to expand from $5.4 billion in 2024 to about $50.3 billion by 2030.
4. How our lives will change: “Excel, email and reports change first”
AI agents and SaaS sound grand, but in practice they are quietly starting to change the simple, repetitive tasks we do every day.
- Personal work
- Automatic sorting, summarizing and drafting of emails
- Automatic meeting notes and to-do lists
- Automating repetitive formatting in Excel and documents
- Inside the company
- HR: automatic resume screening, interview scheduling, onboarding guidance
- Finance: automatic tallying of sales and costs, drafts of monthly reports
- Customer service: AI agents handle simple inquiries and pass only complex cases to people
- Consumer services
- Shopping: agents compare prices and handle refunds and exchange requests for you
- Finance: bundling card, deposit and investment accounts to automatically generate a “this month’s money flow report” every month
- Health care: health-coach agents that automatically adjust exercise and diet based on wearable and hospital records
In the end, much of “what we used to click and write ourselves” will shift to agents handling it quietly in the background.
5. How has it already affected the stock market?
1) “AI agents” have become a new investment theme
- AI chips (GPUs), cloud and data center infrastructure companies: more agents means rapidly rising demand for compute and storage
- Business SaaS companies: re-rated as “agent features” are added to existing SaaS (now seen as work automation platforms)
- Asset management and financial software: money is flowing into AI agent–based investment and risk management solutions
In fact, billions of dollars are rapidly flowing into investment solutions based on autonomous agents and into AI-focused ETFs.
2) Productivity and margin expectations → valuation premium
Companies that use AI agents well are expected
- to grow revenue without adding headcount
- and to raise profit margins by cutting labor and operating costs.
As a result, a premium is already being attached to companies with a story of “AI agent adoption → higher productivity → higher margins and growth.”
3) The market itself is turning into “machine vs. machine”
Algorithmic trading has been around for a long time, but AI agents go beyond simple trading logic: they analyze news, earnings and macro indicators at the same time, adjust their strategies on their own, and even create new ones.
That can make markets more efficient when liquidity is plentiful, but in a crisis it can amplify herding and volatility as “machines all run in the same direction at once.”
6. Where AI agents will hit the stock market hardest: what to watch
- Infrastructure layer
- GPUs and AI accelerators, memory, AI-optimized servers, networking and power infrastructure
- The more agents there are, the more demand for computing, storage and power grows structurally.
- Platform layer (AI agent and orchestration platforms)
- Platforms that bundle multiple agents to run complex work (for example, accounting + tax + financial planning) simultaneously
- These are solutions that let companies build their own “agent teams” and plug them into internal systems.
- Application and SaaS layer
- Depending on how well existing SaaS products such as CRM, collaboration tools, ERP and developer tools build in “AI agent features,”
companies in the same industry could end up either re-rated or left behind.
- Changing business models in finance and investing
- AI agent-based advisory, automatic rebalancing and integrated multi-account management services
- In fund and ETF management, too, how well firms use agents is likely to become a new competitive edge.
BITPRESS Insight
When looking at AI agents and SaaS as an investor, rather than trying to bet on the whole index, first sorting companies into “AI users, AI suppliers, or both” can be a meaningful strategy.
Three groups are likely to be at the center of the medium- to long-term structural gains: 1) companies supplying AI infrastructure (GPUs, memory, cloud), 2) companies providing agent platforms and tools, and 3) companies that embed agents deeply into existing SaaS and services to lift productivity.
In particular, companies that start explaining in earnings releases or IR materials “exactly where and how they are deploying AI agents,” backed by numbers (cost savings, workforce efficiency, new revenue), deserve to be tracked separately as candidates for another round of valuation re-rating.
If SaaS was the rental business of the digital era, AI agents are its staffing agency. As investors, we should not stop at finding AI technology impressive; we should pay attention to how the way companies make money is changing. In the past, platforms with the most users were king, but going forward the market will be ruled by companies that hold the trust and data that let you hand work over to your agent with confidence. Check whether the tech stocks in your portfolio are simply selling tools, or evolving into companies that use those tools to deliver finished results to customers.
