AI is a constantly talked-about topic in conversations in our business and personal lives. It feels huge and fast-changing, and the ramifications can feel uncomfortable. Although there’s been a lot of talk in the press about how AI can reduce costs, there’s been less thought put into how AI can help to produce better quality outcomes.

As experienced web developers, we’ve been considering where, if used responsibly, AI can help to produce superior outcomes in web development that can aid us as human developers. Read on for our thoughts:

Design

AI can be a support in the initial generation of concepts. It can help to provide a wide range of different ideas to help spark a route or direction for a designer, or it can help an experienced designer to explore an idea they’ve already developed, but in a different way.

However, its positive impact on design is limited to these early stages.  AI design lacks emotion; it lacks the human understanding and connection that are critical in building a successful brand – one which a website plays a huge role in.  The all-important emotional connection and resonance needed for humans to believe in something is weak, and the creative outcome feels soulless if it is built only through AI.

AI design beyond initial concepts tends to be generic. AI, after all, is built on existing datasets. There is also the question of copyright and originality to carefully consider.

“As a support at concept stage, AI can be a help, but it isn’t the whole solution and still needs a capable designer to own the process and development of an idea and design to meet the needs of the client and work effectively through a digital platform.”
George, Olamalu’s Lead Web Designer

Technical Development

The best technical solutions are built through a clear understanding of the objectives, technical expertise, and a consistent commitment to building quality and robust security measures.

For now, this isn’t an area that can successfully stand alone through AI. It could be argued that web development can be undertaken through AI and that the process, when compared to that of a human developer, is much faster.  Prototyping, coding and testing can all be undertaken at speed using AI tools. But how does this impact long-term maintenance, and how do we manage the potentially severe data security risks that AI could introduce?

In a world where serious security breaches take place daily on websites across the globe, following robust and systematic practices in all the steps of web development is more important than ever before.

Rapid prototyping using AI tools to generate instant wireframes has the potential to enable quicker decision-making. In theory, it could drastically reduce the time taken for clients to provide feedback and their thoughts and to incorporate those into an updated design. However, good prototyping still requires careful planning and understanding of the needs of the audience. Those needs still have to be turned into requirements. This is more difficult than it seems, as anyone involved in software development projects has probably experienced.  If this isn’t done right, the risk is that the site doesn’t perform or deliver in the way that the client envisaged – meaning that costly fixes may be required later down the line.

AI coding tools promote themselves with the promise of significant productivity. That promise of speed needs to be balanced against the risk of poor quality control.

Solid code written by an experienced professional will be of high quality.  It will be considered and written to perform the tasks agreed at the briefing process and will be scalable, ensuring that the site is future-proof.  Security will have been at the front of the developer’s mind at every step.  What’s more, that code should have been code reviewed by another experienced professional, helping to ensure that nothing’s been missed.

The challenge for software developers using AI to code is how to ensure that the code produced is actually quality code. The most thoughtful approaches here treat AI-generated code as a draft rather than a final product. Effort is then spent on ensuring that the code aligns with the project architecture, covers edge cases and doesn’t introduce any security issues. That’s why AI-generated code must be marked as such so that any reviewer is fully aware of what they’re reviewing.

“From the initial understanding of the brief, to mapping out a solution, building wireframes to support the process, writing robust and secure code, testing, using the insights from testing to tweak and improve, to solid security practices… these are all the steps that a human team needs to undertake to ensure a solid, rational, long-term solution with minimal security risks.”

- Christoph, Olamalu’s Technical Director

Customer Service

Looking after a client for the long term, long after their website has been delivered, is a really important area for us.  We want our clients to get the very best from the site we’ve built for them and offer a range of care packages to help them to do this.

AI can be useful in the organisation of queries; chatbots can help give immediate help or signposting and can sequence requests, track progress and deliver final reporting.  However, the reality for the person on the other end of the process who needs help is that they want to connect with someone, not something, very early in the query process. We don’t want to treat our clients as a ‘ticket’ and firmly believe that in a world that is fast becoming faceless, this is an area which sets us apart from our competitors.  All queries come through our human customer service team. They know every client, understand their business, and this alone can be a big step in helping to solve an issue.

AI can’t get to know an organisation and its people in the way that a human analyst can. This understanding enables the client to receive useful guidance, training and recommendations to help them to thrive.

What does Beverley Smith, Support Manager at Olamalu, think?

“It’s our mission for clients to get the very best from their website.  We want to solve any niggles quickly for them, but it’s more than this.  Our job can be to explain, understand, reassure, train, recommend and fix.  We understand their business and their needs and are their technical support function for the long term. AI can help with the organisation, tracking and reporting of requests, but the real support comes from someone who knows the client’s business.”

Our final thought

There are elements within the web development process where AI could have a legitimate role. However, the need for human involvement and oversight at every stage is non-negotiable in our opinion. Websites support organisations run by people for people, and they need people who understand them and their technical and business needs to build successful solutions for the long term.

We are Olamalu, technical experts and website developers. We’re a friendly and down-to-earth team based in West Oxfordshire, who work together to achieve brilliant outcomes. We’ve been developing websites and designing tailor-made tech solutions for 16 years.