How AI will *actually* transform work
The new roles emerge from the addition of gen AI in our workflows
A couple of weeks ago I wrote about how emerging generative AI tooling is enabling a new wave of software development. You can, if you like, put something decent together with very little coding knowledge using something like Claude’s Artifacts or even just ChatGPT.
I think it’s very easy to get lost in what you can create instantly and on the fly with natural language; it feels so frictionless and exciting that it’s very temping to just play around with some basic ideas and never really take them anywhere. Think about what you can achieve with generative AI as a dabbling solo creative or engineer right now — then think about what you could build when collaborating with a team of other professionals.
Using natural language to build things unlocks so much: it’s not just about building run-of-the-mill software faster — it also enables us to collaborate on making full-blown agentic applications and novel experiences. As AI tooling evolves, and workplaces evolve in tandem, we will absolutely start to see — and kind of already have — new roles emerge.
Through building with Handshake and working across different development teams, I’ve started to identify five new roles or archetypes that represent necessary skill sets for building AI applications. These bridge the gap between traditional software engineering, design, and product management, and are:
☝️ The Technical Guide: this person acts as the liaison between non-technical team members, and the application frame work. They will take natural language prompts, evaluate & test them, and integrate them into system architecture in a way that doesn’t break anything — which means they have a technical understanding of how to build applications, as well as knowing what a good prompt looks like.
✌️ The Hybrid Product Manager: this is a role that incorporates traditional product management skills with light technical knowledge. This person can provide high-level steer of product development, as well as navigate version control via Git, and understand the product’s technical stack. This means they can leverage AI tools to assist with taking a new feature idea or general vision for the product straight into technical implementation.
🤟 The No-Code Ninja: this person is great at whipping up demos and prototypes using no-code tools. AI application development can feel super open-ended — especially to clients who might think the possibilities are endless when they’re not — so it’s vital to be able to put together decent interfaces and functional prototypes to secure buy-in by showing what’s possible. This is also super useful internally, for doing quick tests and exploring different avenues before fully committing to full-scale development.
🖖 The UX Architect: this is an extremely influential role in AI app development. I’ve written before about how generative AI is shifting our UX paradigm away from being command-based to being more intent-based — where the user asks for what they want rather than giving a series of commands. The UX Architect will therefore have the skills to create new types of interfaces and design systems that can capture user intent, and translate them into actionable backend interactions, that are feasibly aligned with what the AI is capable of.
✋ The Natural Language Developer: this role will evolve from what we currently call a prompt engineer. Someone in this role can do a lot more than just successfully interface with a chatbot to get a good output — they are using natural language as if it’s a programming language. They write complex sets of prompts that form the control flows of an application — they are doing exactly what you might do with Python or Javascript, but with natural language.
With these new roles, there’s basically no one working within their own stream: there’s a lot of need for technical and non-technical people to connect with each other regularly. So, with the emergence of these new roles and ways of working together, we’re also seeing — and hopefully will also see more of — the emergence of alternative collaborative tools: visual programming tools like ChainForge, or prompt management platforms like Humanloop are just a couple of examples.
I want to note that these five archetypes are just imagined roles based on what I’ve seen so far — I think a lot of these skills and activities are getting shoe-horned into existing development and project management roles, and obviously these roles will change and adapt as the capability of AI tooling evolves. We’re at a really exciting time now where we suddenly have all the room for innovating on not just in what we make, but how we make it.
There’s so much potential to start collaborating with each other in these fantastic and unique ways right now — I’d love to hear about what kinds of things you’re working on, or what you’re thinking about working on, and which of those five archetypes resonates most with you. Hit reply and let me know! Or leave a comment if you prefer…