Frequently Asked
Questions

You will probably have lots of users with various spreadsheets, Google Docs, file shares etc to support their processes. You have a process that everyone knows by rote but things go awry and some customers get bad service, you take on work that you shouldn’t have or quotes go out wrong. Things feel out of control, users are struggling, customers are complaining.

Yes, a bespoke solution can help to structure and support processes, enforce data consistency, and deliver processing efficiency.
A well-designed bespoke system fits hand-in-hand with your process, making life easy for your employees.

Definitely! Spreadsheets popuping up everywhere usually indicate a process that has evolved without an IT strategy, but they can incur a lot of risks. Building a solution to replace the spreadsheets can really help to properly define and encapsulate that process, removing risks associated with process consistency and data storage.

Yes, usually, because they fit your processes perfectly and evolve with your needs easily. The cost of ownership is lower. Integrations are easier. Users are happier. You own the source code. You control your data.
The perceived benefits of SaaS/Off-the-shelf solutions have largely been negated by AI.

A bespoke developer should be someone that you can view as a long-term partner. They need the basic ability to deliver and support a system for you but it’s also important that they demonstrate an understanding of your business and can help guide you in decisions that affect your solution. In today’s world, they should be leveraging the benefits of AI-powered development because this delivers the best results, fastest, at the lowest cost.

Yes, much! AI works much faster than human developers so it's much cheaper, even compared to offshore development houses. However, we need to consider that there is more work required to give it tasks and manage its work. Overall we reckon that AI-Powered development costs between 10%-20% of the traditional cost to build a bespoke solution.

Yes if the AI is well managed it should be much faster. One AI process can do the work of a human developer but in a fraction of the time, while it is possible to spin up multiple AI agents in parallel to create a team as big as you need, all working together. There are risks with going this fast but with experience, good process and care it's possible to deliver applications in less than 10% of the traditional timescales.

Yes, it used to be a bit random back in 2024 but it has improved immensely since then. We believe that AI actually produces better results than typical human developers, particularly given the inherent speed advantage of AI, which means there is more time for testing whatever gets written. We have AI doing code reviews and testing output in addition to human-in-the-loop quality processes.

As well as being much faster and cheaper than traditional development, AI Powered dev makes it quick and easy to add all the advanced features you need, with minimal cost. This means you can get an even better fit to your business requirements than previously possible. Also, the faster dev time, combined with rapid prototyping means minimal demands placed on your team while your solution is generated.

Yes, they can, but it's the processes that surround the coding (tech design, code analysis, security, deployment etc) that still require technical skills, so the unskilled user still has many challenges to overcome in order to successfully build and deploy a system with AI.

Pretty much all of today's popular languages are available — HTML, JavaScript, C#, Python, Rust, Ruby, Swift, PHP. The AI models are trained on vast datasets that include source code in older languages from years ago such as Cobol, Pascal, and Prolog, so most common languages are available provided they have seen them during their training.

Vibe-coding tends to imply the use of AI to build code without understanding what the AI is doing which can lead to releasing flawed or risky solutions. Here at Provanta we were coding commercial-quality solutions for decades before AI arrived and so we understand how to instruct AI, we know what we are expecting the AI to do and how to deal with what it produces.

Since the advent of AI-Powered coding, yes, bespoke solutions are often cheaper to build and run than the total cost of ownership of the off-the-shelf equivalent.

Not really because bespoke solutions can easily be created and continually evolved to be a perfect fit for your business needs. An off-the-shelf solution may turn out not to fit your needs, today or in the future, and you may not find out until it's too late. You may not be able to adapt the system or you may face a long wait for a solution. Some companies don't like off-the-shelf (SaaS) solutions because they force you to share databases with their other customers. You may also find that a SaaS provider changes or adds features in a way that doesn't suit you. All of which represent risks to a business that relies on such systems.

Bespoke solutions have always been the best option for any business because they fit perfectly. However, bespoke development used to be viewed as time-consuming, expensive and high-risk due to the laborious process of writing code by hand. SInce the maturing of AI-Powered development at the start of 2026 those issues have been eliminated and so today, bespoke solutions are generally the better choice.

Absolutely! We use AI to generate modern versions of solutions from legacy code bases. If you have the source code for the legacy system then we can analyse this with AI to produce technical specifications and then pass this along with carefully crafted prompts and the source code to AI generation tools that will create a new equivalent system on a modern platform ready for another 10 years of production use!

Yes, although its a slightly more complex process. Our tools enable us to understand your old system by analysing binaries or using AI to follow the system screens. We can analyse the output from this to produce technical specifications which direct AI generation tools to create a new equivalent system running on your prefered platform.

The key question here is whether your old system still reflects your current business needs and processes. If it does, modernising it may be the more cost-effective and straighforward option - plus you can always add more features once the new version is running. However, if the system is outdated and no longer aligns with your business goals, building a new solution might be the better approach. Whichever path you choose, AI-powered development will make the process far less painful that your previous exeriences!

It can. If you point your preferred AI (we recommend Claude Sonnet) at your existing code files and ask the right questions then it can give you meaningful insights and helpful answers.

A solution prototype is a preliminary version of a software solution that is built to demonstrate the functionality and user interface before full development begins.

Solution prototypes are useful for demonstrating functionality, gathering feedback, and validating concepts before committing to full development. They allow stakeholders to visualize and interact with the proposed solution early in the development process.

AI has significantly accelerated the prototyping process through automated code generation. This also makes the cost of prototyping negligible, allowing developers to focus more on creative solution design and strategic aspects of the project.

We can build initial prototypes in a few hours or a couple of days. Detailed fully functional prototypes take a few days to a couple of weeks for large/complex requirements. We are talking fully functional systems here that work and look just like the real thing.

Prototyping allows us to validate ideas, gather feedback from stakeholders, and identify potential issues early in the development cycle. This eliminiates suprises later in the project, meaning a smoother delivery process, a better end-product and accurate costings.

A solution prototype is not built for production grade usage — it's not secure, has no ability to share data, its architecture is not scalable. It's built just as a demonstration of how the finished system will work, look and feel. A full solution will have a proper database, decent code structure, security features and will scale to whatever usage level is required.