Interview with Peter Russell: SaaS in 2025 - The Beginning of the End

In the early 2000s, SaaS disrupted the software industry by addressing the cost and burden of building and maintaining a complicated internal IT infrastructure. Instead of hefty upfront investments (CapEx), a significant barrier to entry, businesses could instead subscribe to cloud software solutions, shifting the cost to operational expenses (OpEx) through monthly fees. The 'cloud' model promised scalability and flexibility, but a brand new set of challenges emerged.

The primary issues with SaaS today are its 'pre-packaging' and lack of customisation. In a bid to become giants like Microsoft, SAP or Oracle, SaaS providers raced to develop extensive ecosystems to entrap their customers in, resulting in bloated packages with features irrelevant to many businesses (especially the progressive ones). Same old Jack of all trades, master of none followed by the illusion of choice since you select only from a few pre-defined packages and pay for the whole package even if you only need a fraction of its functionality - behold the next generation version of Shelfware. Customisation options are limited, forcing users to adapt to pre-defined workflows, pester the developers for critical functionalities through product the community or just play along. A far cry from the promised agility resulting in multi-solution ecosystems with overlapping functionalities adding to the confusion of both employees and customers. The promise to take control of data and streamline operations didn't fully materialise, just moved the data silos from offline to 'the cloud'. Allowing SaaS ecosystems to connect and allow data exchange? Forget it! Use Zapier if possible or hire someone to develop you a bespoke API.

However, just as cloud computing disrupted traditional software, AI is poised to disrupt the SaaS landscape in 2025. AI offers a more adaptable approach by providing a more user friendly interface to maximise the data value in a no code environment. Instead of full-stack development, Ai itself, AI-generated tools or even Notion templates deliver immediate results, replacing the need for complex and costly SaaS suites as we know them today.

2025 will push SaaS to being a model for AI to learn specific skills, like Amazon's Alexa, eliminating the need to develop them from scratch and use them through prompts. That will allow businesses and end users customise and refine their AI model to suit their specific needs. In essence, we will be paying for accelerated results rather than recurring subscriptions This shift is fuelled by the progress in AI embedded hardware running entirely new 'operating system'. Powerful and compact AI computers, just like Nvidia's project Digits, capable of matching the processing power of today's data centres sitting comfortably on people's desks. This pushes the need for extensive cloud infrastructure to scaling-up phase, further challenging the traditional SaaS model.

2025 is the dawn of modular, results driven SaaS, where users pay for specific outcomes rather than pre-packaged solutions. This starts a new era of personalisation, but more importantly efficiency, waving goodbye to the limitations of today's SaaS.