There’s a lot of noise about AI right now. Clients are told it’ll make everything cheaper and faster. Agencies are figuring out how to use it without cutting corners. Somewhere between those two stories is the truth, and we’d rather tell you ours plainly than let you guess.
We’re a team of around 25 to 30 people who are genuinely good at what we do, and we’d like you to know how we’re putting AI to work, and just as importantly, how we’re not.
A note on what we mean by “AI”
When we say AI in this article, we mean large language models. The kind of tools behind Claude, ChatGPT, Gemini, and the coding assistants our developers use day-to-day. Not chatbots bolted onto plugins. Not image generators. Not magic. They’re powerful tools that are very good at certain things and unreliable at others, and the difference between using them well and using them badly comes down to judgement.
Where we use AI
We use AI in places where it makes our work better, faster, or more thorough, without changing who is responsible for the result. A few concrete examples.
Ticket creation and structure
When work comes in, AI helps us enrich and structure ticket descriptions so every stakeholder gets the relevant context in a consistent, digestible format. Less back-and-forth, fewer misunderstandings, and a clearer trail of what was asked for and why.
Translating technical work into plain language
A lot of what we do is technical, but most of what you need to understand isn’t. AI helps us turn dense technical explanations into clear, non-jargon summaries so our communication with you is easier to follow without losing accuracy.
Research and discovery
When we’re scoping a new project or investigating an issue, AI helps us move faster through documentation, comparable solutions, and competitive context. The thinking and the decisions are still ours. The lookup work just takes a fraction of the time it used to.
A personal coding assistant for our developers
Each of our developers has AI on hand as a kind of personal assistant. Something to bounce code off, suggest implementations, explore unfamiliar APIs with, or speed up the more mechanical parts of writing software. The result is that our developers work more efficiently and hit scope and estimates more reliably, which translates to delivering on time and on scope more consistently for you.
Test generation and code review
AI helps us generate unit tests and increase test coverage on the code we write. More tests mean more bugs caught before anything reaches your site. It also assists in reviewing code to catch bugs, flag quality issues, and surface potential security concerns early, well before anything ships. It doesn’t replace human review, it adds another layer to it.
Administrative work
The small, repetitive work that used to eat into our day, the formatting and summarising and organising, gets handled faster with AI assistance. That frees us up to focus on the parts of your project that actually need a person thinking hard.
What AI doesn’t do at our agency
This part matters as much as the list above.
It doesn’t replace human creativity, experience, or expertise
The judgement that comes from years of building WordPress sites, managing complex e-commerce platforms, and working with clients across industries is human. AI helps us get to the answer faster. It doesn’t decide what the answer should be.
It doesn’t make the important decisions
AI isn’t truly intellectual. It pattern-matches well, but it doesn’t actually understand your business, your customers, or the context we’ve built up as your partner over months or years of working together. That kind of nuance, the things we know about your brand, your priorities, your constraints, the conversations we’ve had, is what shapes the architectural and directional decisions in your projects. Those decisions stay with us, with the people who know you, and they always will.
It isn’t accountable for anything
Every line of code, every email, every recommendation that involves AI is reviewed and approved by a person. That person is accountable for the result. We don’t shift responsibility to a machine, and we don’t let AI output ship without human eyes on it.
It doesn’t replace our developers, designers, or client success managers
We see AI as something that augments our people, makes them faster, sharpens their work, removes friction from the boring parts. The team you work with is the same team. They just have better tools.
It doesn’t see your data without our control
We’re deliberate about what kind of information goes near which kind of tool, and we treat client data with the same care whether AI is involved or not.
Will AI make my project cheaper?
Honest answer: it depends on what you’re paying for.
The execution parts of our work, like writing boilerplate, generating tests, drafting documentation, structuring tickets, have genuinely sped up. You benefit from that in faster turnarounds and more thorough work, not in us cutting corners on the parts that need human judgement.
The strategic parts haven’t sped up, and they shouldn’t. Understanding your business, designing the right solution, making architectural decisions, reviewing code for security, managing risk. That’s where the value is, and that’s where we still spend the time.
So yes, AI lets us do more in less time on certain things. It doesn’t turn a complex project into a cheap one, and we’d be cautious of anyone telling you it does.
Our policy on AI
We’re conscious about how we use AI, both as individuals and as a company. A few specifics worth knowing.
A written policy every employee reads and acknowledges
Our AI policy defines what AI can be used for, what’s strictly prohibited, and how data must be handled. It’s not optional. Adherence is a condition of employment.
An approved list of AI tools
No employee uses an AI tool for work without it going through a formal review first, covering data security, privacy terms, GDPR compliance, and intellectual property. New tools get vetted before they’re added. Tools that no longer meet the bar come off the list.
A living document, reviewed regularly
AI changes fast. Our policy gets reviewed and updated regularly to keep up with new tools, new risks, and new best practices. The version we operate under today won’t be the version we operate under in a year.
Our overriding concern is your security and the trust you place in us when you let us touch your website, your code, and your business. The policy exists to protect that trust.
AI in Qala
Qala is our product built on top of WordPress, WooCommerce, and open source. A fully functional foundation we use to deliver real, high-value work to our clients faster and more reliably than building from scratch every time.
AI and Qala are a natural fit, and we’re exploring several directions for bringing them together in ways that feel useful and natural for the people actually using the platform. That includes helping you run your site and your business more easily, and making it simpler to understand what Qala can do, and what you can do with it.
We’re early in this work, but we’re moving on it. Expect some genuinely exciting things from this space in the coming months and years.
How clients can, and do, use AI
Some of you are already using AI in interesting ways alongside us, and we’re excited about what’s possible. A few directions we’re seeing or actively exploring with clients.
Content and translation
AI is genuinely useful for drafting content, translating between languages, and adapting tone for different audiences. We’ve helped clients set up workflows that speed up content production without sacrificing the human editorial layer that keeps things on-brand and accurate.
Accessibility
AI can help generate alt text for images, flag accessibility issues, and keep your content compliant. A meaningful help if you’re publishing a lot of content and want accessibility to keep pace.
SEO and content strategy
Using AI to analyse your existing posts, identify gaps, suggest topics, and help draft new content is a workflow that’s working well for some of our clients. The AI suggests, a person decides.
Agentic features inside WordPress itself
With WordPress 7.0 introducing an AI framework into core, we’re going to see AI capabilities native to the platform. We have plans to expand and explore that with clients who are ready to experiment. This is a space we’re particularly interested in.
Conversational access to your business data
Imagine asking your shop, in plain language, “Which products sold best last month compared to the same month from the past two years, and what conclusions can we draw?” and getting a useful answer. That’s the direction things are heading, and we’re already prototyping pieces of it.
If you’ve got an idea about where AI could fit into your site, your shop, or your platform, talk to us. Some of the best uses we’ve found have come from client conversations.
What’s next
Nobody knows for certain what the next few years of AI will look like. Plenty of people are happy to declare their vision of the future and call themselves trailblazers. That’s not really our style. We’d rather be honest: we’ll keep adopting tools that demonstrably help, we’ll keep our policy current, and we’ll keep human judgement at the centre of our work.
What matters to you isn’t the tools we use, the code we write, or our internal processes. It’s the output we deliver, the value it creates, and the fact that you remain a first-class citizen of our business. That part doesn’t change, regardless of what AI does next.