For most of the last two decades, the economics around custom business software were fairly straightforward.
Unless you were a large enterprise with a substantial technology budget, building internal applications usually carried more complexity, cost, and long-term overhead than most organizations wanted to take on. Even companies with very specific operational requirements often defaulted to SaaS platforms simply because custom development was difficult to justify financially.
That shaped how a lot of businesses evolved operationally.
Teams adapted their workflows around the software they purchased. Processes became constrained by platform limitations. Gaps were handled manually. Spreadsheets filled the spaces between disconnected systems. Everyone understood the tradeoff, but for the most part it was still the most practical option available.
What seems to be changing now is not necessarily the need for SaaS platforms. Those are not going away. Most organizations are still not going to build their own accounting software, CRM, or payroll systems.
What does appear to be changing is the viability of smaller, highly targeted operational applications that sit around those platforms and solve very specific business problems.
That shift is being driven partly by AI, but probably not in the way most people think.
The public conversation tends to focus heavily on AI generating code, which is certainly part of the story. In practice though, what feels more significant is the change happening around development workflows themselves. Modern AI-assisted development environments, connected repositories, streamlined deployment pipelines, and cloud-native infrastructure are dramatically reducing the amount of friction involved in building and evolving software.
Historically, even relatively small internal applications could become operationally heavy. Requirements gathering took time. Development cycles stretched out. Small changes turned into backlog items. Enhancements became projects. Deployment carried enough complexity that businesses often delayed improvements simply to avoid disruption.
That friction influenced software decisions for years.
Now the process is becoming much more iterative.
Applications can be prototyped quickly, refined continuously, deployed incrementally, and adjusted alongside the business itself instead of waiting for large future roadmap cycles. Once development and deployment environments become more fluid, the economics around internal software start to look very different than they did even a few years ago.
That is where things get interesting for SMBs.
A lot of businesses have operational workflows that are highly specific to how they function. Quoting processes, onboarding flows, customer communication systems, reporting layers, scheduling coordination, approval chains, inventory visibility, internal AI assistants, service workflows, and operational dashboards are rarely perfect fits inside off-the-shelf platforms.
Historically, companies tolerated those inefficiencies because building custom operational tools often created even larger inefficiencies.
Now there are situations where building the tailored workflow may actually be the more practical option.
Not because custom software suddenly became simple, but because the cost and complexity of iteration have changed so significantly.
That distinction matters because software projects have traditionally been evaluated as large upfront investments. Build everything, launch everything, and hope the requirements remain relatively stable over time.
Increasingly, internal applications are becoming more evolutionary in nature. Businesses can start with a focused operational problem, deploy something relatively lightweight, gather feedback, refine the workflow, and continue iterating over time. In many cases, the first version no longer needs to be perfect because the system itself can evolve much more continuously.
That changes the risk profile around internal software development quite a bit.
There is also a broader strategic implication underneath all of this. For a long time, software increasingly standardized how companies operated. Businesses across industries ended up using largely the same platforms, the same workflows, and often the same operational structures because adapting to software constraints became part of normal business operations.
As tailored operational software becomes more accessible, companies may begin differentiating themselves operationally again. Not necessarily through massive enterprise platforms, but through smaller systems that align more closely with how they actually work.
Some organizations are going to move faster here than others.
The companies that benefit most probably will not be the ones trying to build everything themselves. More likely, they will be the organizations that become better at identifying operational friction points and building focused systems that improve how work actually gets done internally.
That feels like a fairly meaningful shift from where the market has been for a long time.