‘Agentic AI’ is no longer just a buzzword. Growing numbers of businesses are using agents to take autonomous actions across complex environments. These organisations are already unlocking benefits, reducing the number of repeatable tasks for human workers and driving smarter decision-making.
Beyond these immediate benefits, AI agents are also set to usher in a new era of custom-built enterprise software – leading a shift away from traditional SaaS. In fact, a Microsoft executive recently said that AI ‘Business Agents’ will kill SaaS by 2030. The opportunity is huge – with the potential for organisations to create scalable and high-quality software at an unprecedented pace.
Convenience at the cost of customisation
Although SaaS helped organisations to scale quickly, its convenience has always come with a compromise. By design, SaaS platforms are built to serve thousands of customers, relying on generic workflows rather than the unique nuances of individual businesses. As a result, many businesses have become shaped by SaaS software, instead of shaping their software around the business itself.
This means that rather than developing technology designed to play to their strengths, organisations often adapt their internal processes and services to fit SaaS vendor models – standing in the way of differentiation. While teams can tweak settings or add integrations, true customisation – building tools that reflect proprietary knowledge or unique workflows – is largely out of reach.
In an ideal world, every organisation, especially those with large amounts of proprietary knowledge, would build apps tailored to its unique workflows and strengths. However, until now, there have been several barriers to building custom apps.
For one, they’re expensive to maintain. The cost of feature enhancements and bug fixes can push enterprises towards off-the-shelf SaaS alternatives. Building and upholding reliable internal systems also takes a large amount of time and requires specialised expertise.
Organisations may not be able to dedicate so many hours to one single project, or struggle to find and retain the people needed to design, scale and support the solutions they build.
Redefining software with AI agents
Enter agentic AI: some AI agents can help businesses to build custom applications, whilst limiting the historically painful parts of application design. AI agents do so by analysing business processes, orchestrating workflows, and even generating or configuring software, giving teams greater control and flexibility over the applications they create.
What’s more, as these agents operate autonomously, the need for manual coding or specialised engineering resources is reduced. This makes it more realistic for organisations to build applications that are focused on their own workflows and business strengths.
By minimising traditional maintenance inefficiencies, AI agents empower teams to focus on innovation rather than repetitive tasks. This reduces dependency on large maintenance teams, enhances flexibility for seamless feature additions, and accelerates adaptation to business changes.
Applications created with agentic AI are also able to evolve with changing business needs, and remain adaptable over time – ensuring a future-ready, resilient software ecosystem.
Industries with complex, proprietary workflows – such as banking, insurance, healthcare, and aviation – are moving fastest, because organisations in these fields often find SaaS fails to provide sufficient flexibility or differentiation. Take insurance as an example: claims processing, risk assessment, and policy underwriting often rely on specialised workflows.
Agentic AI enables insurers to build apps that reflect these unique processes, rather than conforming to generic SaaS templates. Across healthcare, agentic AI systems can create software that adapts to patient care protocols, regulatory requirements, and research data workflows, instead of trying to retrofit off-the-shelf solutions that were designed for less specialised use cases.
Managing the responsibilities of agentic AI
Leaving SaaS behind means organisations have new challenges to address. With control moving in-house, organisations themselves must take on the responsibility of protecting sensitive data and ensuring applications meet regulatory requirements – all of which can take time.
As models evolve over time, AI-driven applications also need constant monitoring and updates. Without careful oversight, applications can quickly become outdated, no longer aligned with business needs, or prone to errors. This is especially true in cases where AI agents are generating code themselves, as this can easily introduce new security vulnerabilities.
Organisations will need to make sure they have the correct guardrails and a human in the loop to review, test, and maintain code – ensuring ongoing quality and functionality.
The responsibility of AI agents is a lot to manage. Organisations may not have access to the required staffing resource or the appropriate expertise in-house. In these cases, organisations may need to outsource support from a technology partner for expertise on compliance, maintenance practices, and AI governance.
This allows organisations to confidently harness agentic agents, while minimising operational risks.
When custom apps become market differentiators
Businesses that build their own custom applications may even turn them into revenue-generating products, not just internal tools. BlackRock is a standout example. Its Aladdin investment platform began as an internal risk management tool, but went on to be sold to external investment managers worldwide.
This demonstrates how internal innovation, when aligned with business expertise, could one day evolve into a product with global impact – and a new line on the balance sheet.
We might even see the shift away from SaaS happening sooner than Microsoft’s five year prediction suggests. Ultimately, AI agents mean organisations no longer have to be restricted in what they can and can’t offer – they can design, build, and even sell applications designed to support their own employees and their workflows.
What’s more, they can create bespoke tools that their competitors simply can’t replicate. Differentiating themselves in the market in this way will only help businesses in outpacing their competition in a future that will be defined by AI.
