Lately, I can’t stop thinking about the potential of vertical AI agents. It’s a topic I’m really excited about, especially when I compare it to the rise of SaaS we saw over the last two decades. And I’m not the only one; many people in the startup community are beginning to realize just how big this could be.
To put it into perspective, the SaaS explosion happened when technology enabled software to be used over the web rather than installing it from a CD. This shift, largely driven by the XML HTTP request function in web browsers in 2004, led to the creation of web applications that looked like desktop software, such as Google Maps and Gmail, and changed the software landscape.
Now, large language models (LLMs) are generating a similar shift. We’re at the forefront of a new computing paradigm that opens up entirely new possibilities. As with the rise of SaaS, questions arise about where the value and opportunities for startups will be in this new landscape.
Parallels and Differences
Reflecting on the emergence of SaaS, we see three main categories of companies:
Obvious consumer product ideas: Here, the big players like Google and Facebook captured most of the value.
Non-obvious mass-consumer ideas: Startups like Uber, Airbnb, and Coinbase succeeded in areas where the giants hadn’t ventured.
B2B SaaS companies: This category produced the largest number of billion-dollar companies, mainly because there isn’t one “Microsoft of SaaS” that covers everything.
Vertical AI agents might follow a similar path. The obvious mass-consumer applications, such as voice assistants, will probably be dominated by the tech giants. However, the opportunities for startups lie in more specific niches and in automating enterprise tasks.
Interestingly, unlike the early days of SaaS—when mass-consumer applications (email, chat, etc.) led the way—vertical AI agents seem to be gaining traction in the enterprise sector from the start. This could be because there’s already a business culture that values specialized solutions.
The Potential of Vertical AI Agents
Why so much excitement about vertical AI agents? Here are a few key points:
Automation of repetitive tasks: AI agents can automate repetitive and boring administrative tasks.
Replacing entire teams: There’s talk that vertical AI agents could replace entire functions and teams within companies.
Enhanced user experiences: By focusing on specific verticals, these agents can offer user experiences far more user-friendly than traditional enterprise software solutions.
Reduced staffing needs: Companies can cut costs by automating many tasks and needing fewer people. This means companies can be more efficient and potentially have fewer employees—leading to “unicorn” businesses with as few as 10 people.
Huge market: Every unicorn SaaS company has a potential equivalent in the world of vertical AI. AI agents not only replace SaaS software but also many tasks previously handled by employees.
Examples and Use Cases
Some examples of companies already exploring this space:
Outset: Uses LLM to analyze surveys and the Qualtrics space.
Mtic: An AI agent that handles QA testing, eliminating the need for a dedicated QA team.
A Priora: Manages the entire technical screening process for recruiters, reducing the need for traditional recruiting teams.
Cap.AI: A chatbot for developers that answers technical questions, reducing the need for large support teams.
Parel: A customer service agent that can handle complex tasks, reducing the need for large support teams.
GigML: A specialized agent for the instant delivery market that manages customer support tickets.
Salient: Uses AI-driven voice calls to automate collections in the automotive industry.
These are just a few examples, but the potential is enormous. Every vertical has its own needs, and AI agents can adapt to deliver highly customized solutions.
The Future
We’re at a very early stage of this revolution, and the landscape is evolving rapidly. What seems impossible today could be tomorrow’s reality. LLMs are constantly improving, and competition among different models is creating a more fertile ecosystem.
I believe we’re facing a radical change in how we work, and vertical AI agents will be a fundamental part of this future. Instead of fearing AI, I think we should embrace it to build more efficient companies and to improve people’s lives.
What do you think? Are you seeing the same potential in this area? I’d love to hear your thoughts.
Comments