Vacancy Poster is one of many vertical SaaS companies waking up to the smell of fresh coffee: the future may lie less in providing users with another graphical interface and more in exposing specialist capabilities through APIs.
This shift is not being caused by AI alone. APIs have enabled software integration for decades. However, AI—and particularly the emergence of AI agents—is dramatically accelerating it.
Large language models are trained on enormous quantities of historical data. Once a model has been trained, its underlying knowledge does not continuously update each time it encounters something new. An AI model may be excellent at explaining a concept, drafting an article or making your cat appear to talk, but its training alone cannot tell it what happened five minutes ago, whether an invoice has been paid or which jobs are currently available.
To interact with the live world, AI systems need tools.
A tool might perform a web search, retrieve today’s date, check a database, obtain a customer record, calculate an insurance quotation or place a job advertisement. The model can examine the information returned by one tool, decide what it needs next and continue until it has enough information to complete the task.
This iterative combination of reasoning and action is one of the foundations of agentic AI.
Importantly, the human does not necessarily specify every tool or every step. Within the permissions and controls established by the application, the AI agent can decide which available capabilities it needs to use.
That changes the strategic value of an API.
Imagine asking an AI assistant to find you a job.
The model’s original training data cannot contain vacancies posted this morning, so it must obtain current information. It could search the general web, but that is a relatively crude approach. A human jobseeker would be more likely to visit a specialist platform such as Indeed, enter detailed criteria and receive a structured shortlist.
Now imagine that a company created an API that searched numerous job-distribution platforms, normalised their results, removed duplicates and returned a consistently structured list of suitable vacancies.
If AI platforms and enterprise agents adopted that API as their preferred job-search tool, the provider could become an important gateway between candidates and the recruitment market.
It might provide the service to AI platforms at little or no direct cost. The resulting intelligence—what people search for, where demand exists, which vacancies attract attention and where skills shortages are emerging—could itself become extremely valuable.
The commercial prize is therefore not limited to transaction fees. It may include market intelligence, referral revenue, premium placement, analytics and access to demand at the precise moment that a user expresses an intention.
The same opportunity exists in almost every vertical market: travel, recruitment, insurance, property, healthcare, logistics, legal services, finance and procurement.
The next major contest may not simply be to appear at the top of a conventional search-results page. It may be to become the trusted tool an AI agent chooses when it needs a particular capability.
Services such as ChatGPT and Claude are increasingly becoming the place where people begin a task rather than merely search for information.
That does not necessarily mean that conventional search engines will disappear. However, the interface is changing. Instead of receiving a list of links and completing the work themselves, users increasingly expect an AI system to research, compare, decide and act on their behalf.
In the traditional internet economy, companies competed for clicks.
In the agentic economy, they may compete to be called.
This creates both an opportunity and a threat for SaaS providers.
A SaaS company whose value is trapped behind a human-operated interface risks becoming difficult for agents to use. Its carefully designed screens may become less important if customers increasingly interact with its capabilities through an AI assistant, an internal agent or another software platform.
By contrast, a SaaS provider with a secure, well-documented and machine-readable API can make its expertise available wherever the customer chooses to work.
The API becomes the product. The graphical interface becomes one possible client of that product.
As more SaaS companies expose their specialist capabilities through APIs, organisations gain much greater freedom to assemble the best services for each part of a business process.
A company might combine identity verification from one provider, document generation from another, digital signing from a third and specialist risk analysis from a fourth.
APIs make this technically possible. They do not, by themselves, make the resulting process manageable.
Something still needs to decide:
This is where orchestration becomes important.
Traditional middleware has generally concentrated on moving and transforming data between systems. The human-facing workflow often remained inside the graphical interfaces of the individual SaaS or on-premises products.
AI introduces a new participant: an intelligent but non-deterministic system that may consume information, make decisions, generate new data and invoke further services.
Connecting that safely into a real business process requires more than simple data transfer.
A modern orchestration platform must coordinate APIs, AI models, business rules, human tasks, permissions, exceptions and user interfaces. Products such as Opzaro, Camunda, FlowForma, and Process Maker, to name a few, are positioned within this developing space, enabling organisations to combine external capabilities into governed end-to-end processes.
The objective is not to allow an AI model to operate without limits. It is to place AI reasoning inside a controlled process in which permissions, responsibilities, escalation routes and audit trails are clearly defined.
The future enterprise software stack may increasingly consist of four layers.
Delivered through APIs by vertical SaaS providers.
Able to interpret objectives and decide which capabilities to invoke.
Controlling the process, permissions, sequencing, exceptions and human involvement.
Generated or assembled around the needs of the task rather than dictated by each underlying supplier.
This model could weaken the traditional advantage of the all-encompassing software suite.
Instead of selecting one large platform and adapting the business to its limitations, organisations may choose the strongest available capability for each requirement and use orchestration to make those capabilities operate as a coherent whole.
There will still be graphical interfaces. Humans will continue to need dashboards, forms, approvals and oversight. But those interfaces may increasingly sit above the process rather than define it.
The SaaS industry is not merely adding AI features to existing software.
It is beginning to confront a more fundamental change: the next important user of a software product may not be a human being. It may be an AI agent acting for one.
That means SaaS companies must think beyond exposing their existing screens through an API. They need to make their underlying expertise discoverable, understandable, secure and usable by machines.
The winners may be the businesses whose capabilities are easiest to trust, combine and orchestrate—not simply those with the most attractive standalone interface.
In the age of AI, software companies will still need to win users.
But increasingly, they will also need to win the agents acting on those users’ behalf.
Drawing on more than 20 years of experience delivering SaaS solutions to over 800 customers, the product team at Vacancy Poster has created Opzaro: a modern process orchestration platform designed to help organisations of all sizes formalise, improve and automate their business processes.
Opzaro enables businesses to integrate best-of-breed APIs, coordinate human and AI-driven activities, and create governed end-to-end workflows without becoming dependent on a single software suite.
The platform has already been selected to provide the foundation for a major process transformation programme at a global Lloyd’s of London insurance broker with approximately £50 million in annual revenue, following an evaluation of Opzaro alongside established workflow and process-management platforms.
By providing a clear path from today’s processes to tomorrow’s AI-enabled operations, Opzaro helps organisations pursue improvements in quality, cost and speed of delivery while retaining control, oversight and auditability.
Opzaro coordinates APIs, AI models, business rules and human tasks inside one governed, auditable workflow engine.
Explore AI & automationWe'll show you how Opzaro combines your own APIs, AI and people into one governed workflow — not just slides.