SaaS Will Change in 2025: Here’s How
by Phil Edmondson-Jones, Partner at Oxx.
Guest post by Phil Edmondson-Jones, Partner at Oxx.
The global SaaS market will change dramatically in 2025. With increased adoption of AI tools, gaps to fill in underserved markets, and an increased focus on security and regulation, 2025 should see some exciting developments throughout the industry.
Increased investment in ‘niche’ vertical SaaS tools for under-digitised sectors
For a number of years we’ve witnessed a continued trend of growth in vertical SaaS. Certain industries have received more focus than others, for example, insurance, healthcare, construction and the hospitality sectors have received considerable funding to scale dedicated tools to improve the industry’s technology stack.
Subsequently, I predict that 2025 will see more solutions developed for underserved sectors: for example, beauty & wellness, the public sector, veterinary clinics etc -- businesses who still have a material offline (or semi-online) component to their workflows and heavy ongoing use of legacy systems.
SaaS challengers will be able take advantage of this gap in the market to create purpose-built platforms which serve these smaller businesses effectively. These platforms are now cheaper than ever to build using automation, and companies now benefit from a playbook on extending their revenue model established by the vanguard in vertical SaaS, enabling them to build and scale in a much more efficient manner.
As a result, some sectors that had a seemingly impossible cost to serve are now in play, and we’ll see some rapid digital transformation in these sectors.
Magic-feeling ‘agentic-AI’: bubble or the real deal?
‘Agentic AI’ is now in vogue, set against a backdrop of an ongoing media-debate as to whether this is the next bubble, or whether real sustainable value will emerge. Personally, I believe that it will fall somewhere in between -- as usual some will ultimately fail (the products will be exposed as lightweight wrappers without much defensibility), but there is undoubtedly already a wave of young companies doing genuinely interesting and super high-potential work.
One of the interesting early themes for agent adoption in 2025 is the development of agents working on tasks that people generally loathe -- for example compliance / tax / accounting workflows, heavy legwork in outbound sales, legal information gathering, customer or IT support.
Agents are a major breakthrough versus previous automation capabilities, in that they aren’t simple rules based engines, they are capable of evaluating their environment and carrying out complex processes. The best ones already feel like they are already using something akin to ‘judgement’ to complete ambiguous tasks quite effectively.
Users (not just developers) can instruct an AI agent to follow a set of tasks, replacing basic human drudgery with a digital worker, and transforming the way people do business. This will start to have a dramatic impact on human jobs - initially, managing the division of tasks between human and agentic workers will have interesting organisational design consequences.
The landgrab in data & security concerns
There’s a lot of focus on the ‘magic of AI’, but behind the scenes comes the necessary -- and very messy -- task of getting data into shape. This is essential to making sure the data is reliable, and reduces the risk of the AI hallucinating or producing garbage answers. It involves a lot of grunt work: database cleaning, organisation, and downstream process optimisation through developer tooling to make the data into something usable.
At the same time, security and risk management is extremely important to this entire workflow. For all the positives of widespread availability of generative AI, there is a proliferation in cyber attacks as they become cheaper and easier to deploy. Any upgrade to the data and application production stack must have the dual purpose of being high efficacy, but also having strong security credentials.
Defence tech noise: and why supply chain and manufacturing tech could benefit
Right now there’s a lot of noise around defence tech, given the unpredictable and tumultuous global geopolitical stage. This is particularly true in Europe, who have relatively underspent in recent years. There’s a lot of early stage interest in levelling up Europe’s capabilities here, supported by both the public and private sectors.
Rising tensions and the move towards more protectionist policy is also increasing the level of focus on monitoring and optimising supply chains, and driving greater (domestic) manufacturing throughput / reducing downtime. The addition of the financial risks associated with tariffs, on top of the existing environmental and operational concerns has made performance in these areas more business critical than ever.
Companies want to be able to more deeply understand and adjust individual elements of their supply chain to drive more efficiencies and better outcomes, and manufacturers are under intense pressure to modernise given tough trading conditions & thin margins. Similarly both of these areas are seeing a wave of early stage momentum and disruptive propositions are rapidly gaining share.
The swift and cheap scaling of SaaS companies
There have been a few dozen European and US SaaS businesses that have recently made public they have scaled to ~$20m ARR in a matter of months (eg Lovable), and $100m ARR in no more than a year or two (eg Cursor). Many of them have managed to do this without expanding their headcount nearly as much as their predecessors.
Many of these SaaS businesses sell fantastic novel products, with little / no integration or onboarding requirements, that have pretty near immediate time to value. They can be explored and consumed directly by the end-user, at limited cost, and the communities through which they were discovered are very efficient at sharing these breakthrough tools with peers across the world.
I think 2025 will see many more companies come out publicly to show off their exceptional revenue growth - which ultimately comes off the back of the incredible traction caused by millions of people globally experimenting with a raft of fascinating AI tooling, we’re just starting to understand the true capabilities of. But expect more commentary and analysis on retention metrics and cohort behaviour as these become clear. And the trend towards building SaaS companies efficiently, and certainly with fewer employees, is definitely here to stay.
These are just a few examples of changes we could see to SaaS in 2025, and I am excited to see developments in the field and the future of this technology.
