When many people think of AI and business, they have a vision of everyday tasks being undertaken by AI agents. Whilst this will happen, there is another area that many people are not fully aware of.
Bespoke software written using AI.
Traditionally, if you required custom software you had to employ the services of a software development company. This made software very expensive, which is partly why SaaS (Software as a Service) products became popular. Such software is cloud-based, and you pay £x per month or per year for each licensed user.
With the rapid advances in AI, products like Claude Code, ChatGPT (formerly Codex) and Gemini CLI have evolved.
These products, with the right knowledge, can write software for you across many programming languages.
To many, this might not seem that important.
However, for larger surveying organisations, it means they can create their own software that meets their usability requirements and, more importantly, complies with their governance policy. This means feedback from users can quickly be fed back into the software development team to ensure new versions can be created.
Also, building your own software in-house means integration between products is simpler, and aspects like security and usability are more controllable.
With an in-house software team, developers can work alongside domain experts from within the organisation to ensure the products are fit for purpose. This means production time from initial idea to minimum viable product can be days instead of months. Those experts are engaged with the design brief for the AI coding software and are on hand to test it immediately.
For medium and smaller firms, employing a specialist team to produce bespoke software is, at the moment, cost-prohibitive.
Although, going forward, I can foresee software development following the same path as website development. In the early days of website design, you had to learn HTML and write code by hand in a text editor. Along came products like HotDog, which made coding in HTML easier.
Then software like Dreamweaver came along, and eventually drag-and-drop tools that let you build a website without any real understanding of how HTML actually works.
Within the near future, software that helps you write the brief to give to, say, Claude Code will be commonplace, and even small firms will be able to create bespoke software without understanding how it works.
Here at Survasai, we use a domain AI-augmented developer (a hybrid between a domain expert and an AI-augmented developer) to build and test a number of software products for a variety of applications within the building surveying profession, including observation, reporting, and training/education software.