Why AI app-building makes niche software viable for small firms

AI-built apps may be creating software where none existed

One in three operators interviewed by Emergent said the application they built replaced nothing. Emergent says no suitable product existed at a price they could afford. That points to a potentially consequential use of AI app-building: creating software that previously made no economic sense to build.

The finding has a strict boundary. San Francisco-based Emergent interviewed 300 operators who had already chosen its platform and analyzed more than 50,000 live applications. Emergent sells the platform these operators selected, so it has a commercial interest in evidence that its technology enables previously uneconomic applications. This selected population cannot establish how widespread the problem is among small businesses. It can show why some businesses that choose Emergent say they are building software.

Small firms can fall into the gap between SaaS and custom development

A business with a highly specific workflow can face poor conventional options. A packaged SaaS product may fit only part of the process, while custom software may fit closely but cost more than the problem can justify. Continuing with a manual process can then be economically rational.

Among Emergent respondents who had sought development-agency quotes, the reported median price for a basic custom build was USD $20,000. Some reported quotes as high as USD $100,000, with delivery estimates ranging from several months to more than a year. Emergent says cost was the barrier cited most often and that many owners abandoned commissioning plans.

The alternatives reported by these selected builders show how the constraint played out. Only one in seven interviewed operators felt well served by existing software. Among the rest:

Reported situation	Share
No available approach solved their problem	33 per cent
Relied on manual processes	27 per cent
Used poorly fitting software	27 per cent
Had paid for custom development	12 per cent

For these Emergent users, purpose-built software could deliver real value while remaining below the threshold that justified conventional development.

AI app-building changes the economics of missing software

Emergent reports several outcomes from applications built by its respondents:

Outcome	Share
Replaced a software subscription	33 per cent
Retired a manual process	21 per cent
Reduced dependence on developers	13 per cent

Mukund Jha, Co-Founder and Chief Executive at Emergent, said the company had expected subscription replacement to dominate. Instead, he said a third were addressing problems where the market was “too small or their problem too specific for a vendor to chase.” Jha benefits commercially if businesses view Emergent’s platform as a practical way to serve such workflows.

This describes a distinct economic category. A workflow can matter enough to one healthcare practice, retailer, construction business or poultry farm to warrant dedicated software while offering too little total demand for a SaaS vendor to build a product for it. If AI app-building lowers development costs enough, the business closest to that workflow may be able to create its own system.

Emergent also reports signs that some applications perform operational work:

Application characteristic	Share
Takes payments directly	One in four
Sends transactional email	One in three
Uses a custom domain	One in three

Emergent says the applications are used by roughly eight million people across 28 industries. These measures do not establish business value. They show what Emergent counts as evidence that some applications have moved beyond prototypes.

Software creation can move closer to the workflow

For executives, lower development costs can change the calculation behind a software investment. A specific workflow that once remained manual or depended on an imperfect subscription may justify a narrowly tailored internal or customer-facing application when the expected cost, delay or lost revenue exceeds the cost of building and operating it.

For SaaS leaders, the same economics could create competition from applications built inside customer organizations. They could also allow businesses to create software for narrow workflows that packaged-software vendors have little incentive to address.

Retool reports a related development in a different population. Its research found that 35 per cent of enterprises had replaced at least one software subscription with something built in-house, while 60 per cent said they had shipped software without IT sign-off during the previous year. Retool sells software-development tools, giving it a commercial interest in increased in-house development. Its enterprise research measures different behavior among different users from Emergent’s selected operators. The two studies should therefore be read separately.

Easier building leaves a use-case challenge

Lower development barriers still require a business to identify a problem worth solving and understand the process well enough to specify what the software should do.

Research by the British Chambers of Commerce and Atos found that 54 per cent of UK firms were actively using AI, up from 35 per cent a year earlier. Yet 60 per cent cited limited AI skills as a barrier, and 71 per cent said they could not identify a clear use case.

Those findings concern AI adoption broadly and do not establish the barriers to AI app-building specifically. But they identify constraints that executives evaluating these tools may need to test inside their own organizations. Software can become cheaper to create while process knowledge, governance and expected return on investment remain separate management questions.

The potential population is large. World Bank data cited by Emergent puts small and medium-sized businesses at about 90 per cent of businesses globally and more than half of employment. Emergent’s selected-user research cannot establish how common unmet software needs are across that population. Representative research across businesses that have and have not tried AI app-building tools would be needed to estimate how many valuable workflows remain manual, depend on poorly fitting software, or lack dedicated software because conventional development cannot be justified.

Main highlights

New software demand may be emerging: One in three surveyed Emergent users said their application replaced nothing, suggesting AI app-building can make software viable for niche workflows that previously could not justify development.
Small firms face a gap between SaaS and custom development: Packaged software may fit poorly, while agency-built software can be too expensive. Leaders should reassess manual or poorly served workflows as development costs fall.
AI app-building changes the build-versus-buy calculation: Emergent users reported replacing subscriptions, manual processes and developer work with purpose-built applications. Evaluate these tools where a workflow has clear value but limited appeal to conventional software vendors.
Software creation can move closer to business teams: Lower barriers may allow teams with direct process knowledge to build applications themselves. Executives should compare expected operational value with development, maintenance and governance costs.
Easier development does not solve the use-case problem: Many businesses still struggle to identify clear AI use cases or lack relevant skills. Prioritize well-understood workflows and measurable business outcomes rather than building applications simply because development is easier.