Leaf is a tech business who work closely with some of the world’s most recognisable eCommerce brands, writes Helga Alvarez, CTO, Leaf.
The business was founded by a group of Costa Rican software engineers in Newcastle. Although Leaf would now be described as “early adopters” of AI, we never really thought about it that way. We were just doing what software engineers do – using technology to solve problems and work smarter.
Practical tips on how to use AI
With AI now used right across our business, I’d like to offer some practical tips on how AI – especially generative AI – can help tech teams work more efficiently. With lots of the advice out there on how people who aren’t from a tech background can use AI, I’d like to focus on how those already in tech roles can use generative AI to boost their skills and productivity today.
First and foremost, you need to be curious when it comes to AI. There are a multitude of generative AI tools available so don’t be afraid to experiment with them. AI is a fantastic way to augment the skills of our team – we always say it transforms them into superheroes! We encourage them to test a variety of tools and report back to colleagues on what they have discovered.
Navigate with Github Co-pilot
Of all the AI tools we use, Github Co-pilot is arguably the most important. Co-pilot uses an AI codex to suggest code and functions in real time. Today Co-pilot is responsible for around 60% of our code. As the name suggests, the tool doesn’t replace skilled programmers, instead it radically speeds up what would be lengthy and time-consuming processes.
ChatGPT is now completely ingrained in our business. We use it for a variety of purposes such as writing test cases for quality assurances. Again, the real value is not that AI is doing something we couldn’t do ourselves, it is the speed with which it can deliver. For example, we recently needed to map data from one Shopify schema to another – a process which would have taken the team around three days. By using ChatGPT, we were able to complete the task in a matter of hours. Although ChatGPT can offer code assistance and language and syntax explanations, our team’s use of ChatGPT extends beyond technical assistance – some of our team even use AI tools to transcribe meetings and then use ChatGPT to create a summary that they can refer back to.
Introducing an AI gamechanger
The strength of AI isn’t just reserved to your core skills however. Traditionally coding and artwork design have been seen as two separate disciplines. Adobe Firefly has been a game-changer when it comes to allowing anyone on the team to quickly creating images and artwork freeing up our designers to work on more complex tasks.
Sticking with image generation, perhaps one of the most interesting applications of AI is using one tool to maximise the effectiveness of another. Midjourney is a powerful AI image generator however it’s not always intuitive to use. Our solution is to use ChatGPT to craft the right prompts to use Midjourney effectively. Although I’ve picked out Midjourney here, you can apply the principle to any tool you are struggling to get effective results from. Using one AI tool to feed another removes hours of trial and error and helps you improve your skills without lengthy training courses.
The ultimate learning tool
What sets ChatGPT and other similar large language models apart is their ability to “remember” previous responses. This allows you to have detailed conversations on technical questions. I recently used ChatGPT to help develop a complex diagram in Mermaid. Although ChatGPT couldn’t generate the correct prompts instantly, it was able to continually re-iterate based on our interactions until we arrived at the right result.
Enjoy doing the human things AI can only dream of!
So go ahead and experiment. Think of generative AI as a skilled colleague who you can bounce ideas off of, offer you tips, guidance and productivity hacks, and free you up to do the parts of your job only a human can do.
