Horizontal AI
Horizontal AI is AI built to work across many industries and use cases with the same general capability, such as a generic chatbot platform or a general-purpose language model, rather than being built for one sector's specific workflows.
It is the AI equivalent of horizontal software: broad, industry-agnostic, and configurable by the buyer to fit their use case. The trade-off is depth. A horizontal AI platform has to be adapted, through prompts, integrations, and internal expertise, to handle sector-specific requirements like FCA complaint handling rules or KYC procedures, because those were not built into the product from the start.
Horizontal AI vs vertical AI: what's the difference?
Vertical AI is built for one industry's specific problems, with the workflows, compliance logic, and integrations that industry needs designed in from the start. Horizontal AI is built to be broadly applicable, then configured by each buyer for their own use case. Neither is universally better: horizontal AI often reaches production faster for a generic use case, while vertical AI typically needs less custom work to meet a specific industry's requirements, because that work has already been done once for every customer in the category.
Where horizontal AI falls short in regulated financial services
A horizontal AI platform does not arrive knowing what a Consumer Duty vulnerability flag is, how a chargeback reason code maps to a dispute outcome, or what an FCA-compliant escalation path looks like. A buyer in financial services has to build or buy that knowledge separately, then maintain it as regulation changes. That is the work a vertical AI vendor has typically already done for every customer in the category, which is the core argument for choosing vertical AI in a regulated industry over adapting a horizontal platform. We set out what a horizontal platform has to be taught about financial services, and what a sector-built one already knows.
Frequently asked questions
Is a horizontal AI platform always cheaper than a vertical one?
Not necessarily. The licence cost may be lower, but the implementation, integration, and ongoing compliance work a buyer has to do themselves often costs more over time than a vertical platform that already includes it.
Can a horizontal AI platform be made to work in financial services?
Yes, with enough custom configuration, integration work, and in-house compliance expertise. The question is whether that investment costs less than buying a platform built for the sector already, which depends on the buyer's scale and timeline.
Do horizontal AI vendors ever move into vertical AI?
Some do, typically by building industry-specific modules or partnering with sector specialists. Whether that reaches the same depth as a vendor built vertical from the start is a comparison a buyer should test directly rather than assume.