The Death of SaaS as We Know It? Will AI Agents Eat Software?

Remember when we used to install software from CDs? Then SaaS arrived and changed everything. Next, we’ll witness an even bigger revolution: AI agents transforming how business SaaS applications work. While Salesforce, Workday, and countless other platforms transformed how we work, their application layer could be replaced by intelligent agents that bypass interfaces altogether to work directly with you and your data. The applications will change dramatically, but your business data remains the constant foundation of value. Here’s why this seismic shift will reshape every industry and why you need to prepare.

What is it?

SaaS (Software as a Service) transformed business technology by moving from installed software to cloud-based subscriptions. Rather than managing servers and updates, companies simply access applications through browsers and pay monthly fees. SaaS brought faster implementation, reduced IT overhead, and lower upfront costs.

AI Agents represent the next evolutionary leap. These intelligent systems can:

Understand natural language requests (“Show me which customers are at risk of churning”).
Write code on demand to fulfill specific business needs.
Access and manipulate your data without predefined interfaces.
Create custom workflows that adapt to your specific requirements.

For example, instead of navigating through multiple application screens to generate a specific report, you might simply ask, “Which enterprise customers in the healthcare sector haven’t upgraded in the last 18 months?” and receive an immediate answer.

The fundamental insight: While SaaS applications come and go, your business data – customer records and transactions remain the core, enduring asset. SaaS applications are just interfaces to this data. AI agents offer a more direct, natural way to leverage this data asset.

What does it mean from a business perspective?

As SaaS evolves and application layers get replaced by AI agents capable of interrogating data and writing code on demand, significant business implications emerge:

Significantly reduced training overhead – With AI agents, employees simply ask questions in natural language, reducing training costs dramatically.
Hyper-personalisation by default – Instead of all salespeople using identical views, AI agents create personalised views based on their specific selling style and customer portfolio.
Radical efficiency gains – Instead of spending time navigating software interfaces. AI agents could reclaim this time through direct, conversational interaction with data.
Democratised data access – Marketing teams can request “Show me which email campaigns generated the most sales last quarter” without SQL knowledge or waiting for BI team support.
Business-first processes – Rather than conforming to predefined approval flows, AI agents implement your exact desired process, even as it evolves over time (a change in your business processes will not be constrained by your ability to implement it in software).
Seamless system orchestration – AI agents can pull customer data from your CRM, financial history from your ERP, and support tickets from your help desk, all in response to a single query without complex integration projects. (Are AI Agents The New API’s?)
Strategic vendor independence – Your business strategy won’t be constrained by what your SaaS vendors can provide – AI agents can build custom capabilities directly atop your data. (There will always be complexities in the underlying data models that mean the AI agents will likely come from the SaaS companies directly.)
Value-based economics – Instead of per-seat SaaS licenses regardless of usage, AI agents will be priced differently, possibly based on business outcomes and actual computational resources consumed.

What do I do with it?

This transition won’t happen overnight – it is a major shift that will take time, but forward-thinking leaders should begin preparing now:

Data quality – AI agents can only be as good as the data they access. Do you have a data management function – if not, start one.
Vendor Plans: What are you vendors road maps for agent adoption? How does this road map fit in with your own road map for business development – work with your EA team.
Launch an AI pilot program – Start with something like Microsoft Copilot Studio to get started (Copilot Studio will give you a sense of what will be possible with things like Knowledge Sources, Triggers and Actions before going deeper into more powerful tools – check out tools like Langflow.)
Upskill your team – Train your IT teams on developing AI agents. Identify 5-10 “AI champions” across departments and invest in their training on effective prompt engineering and AI collaboration techniques. These individuals will become your internal change agents.
Redefine vendor evaluations – Update your technology assessment criteria to ensure vendor selections take into account their AI and AI Agent road map.

This isn’t science fiction – it’s already beginning but SaaS companies with massive code bases and complex applications take time to change. The SaaS vendors themselves see this coming – see Matthew Berman‘s analysis of Satya Nadella’s video in the link below.

Are you preparing your organisation for the post-classic SaaS era, or will you be caught maintaining expensive legacy interfaces while your competitors leap ahead with AI agents?

What are you doing with AI agents, I’d love to hear about your experiences?

What is it?

What does it mean from a business perspective?

What do I do with it?

Further Reading (and watching)