How to understand technical architecture with AI
🧠Knowledge Series #63: A guide for PMs, designers and non-engineers. Practical ways to get up to speed with technical architecture - quickly.
🔒The Knowledge Series is available for paid subscribers. Get full ongoing access to 60+ explainers and tutorials to grow your technical knowledge at work. New guides added every month.
When you join a new company, one of the first things you have to get to grips with is the technical architecture of the product you're working on. And it's not always easy. In some companies, architectural diagrams are well documented and kept up to date, but in others, the developers who originally worked on a feature may have long since moved on and, often by design, there is no formal process for documenting and storing architectural documentation.
In this scenario, it’s often up to you as the new starter to figure your way around the relevant parts of the product, speaking to each of the various engineers who do know how things work to bring you up to speed.
AI can help fast-track this process in ways that were unimaginable previously. Not only can it help you to understand key technical architecture quickly, but it can also simulate how any potential new features might impact existing architecture.
In this Knowledge Series, we’ll explore some of the ways you can use AI to understand existing technical architecture quickly, visualize technical systems and data flows and even generate your own personalised technical architecture documentation that you can refer back to when creating new features or crafting your strategy so that you don’t end up committing to building stuff that’s undeliverable.
Coming up:
Practical examples - how to use AI to:
Translate complex architectural diagrams into plain English
Use the "personification technique" that transforms complex systems into characters you'll actually remember
Simulate future scenarios when building new features
Understand architectural constraints and prevent tech debt
Generate architectural diagrams you can share in seconds
The new AI tool that can decode technical architecture in minutes – what used to take weeks of engineer meetings
The full list of tools you can use
Why an understanding of technical architecture is useful
It’s certainly possible to work at a tech company and have very limited exposure to the underlying architecture. But there inevitably comes a point where it’s unavoidable; typically, this is when a new feature is written off as too complex or a strategic decision is undermined by tech debt that make it extremely difficult.
For non-engineering members of product teams like PMs and designers, having an understanding of the most important parts of the architecture of the product you're working on can be super helpful.
It can help you to roughly estimate ideas before deciding to progress further. It can facilitate better conversations with engineering teams and help you to understand potential technical constraints.
How AI is changing the relationship between PMs, designers and what you need to know about technical architecture
It’s true to say that AI is abstracting the process of product development away from a line by line analysis of hand written code and that understanding every part of a product’s codebase isn’t really necessary. In the future, it’s likely that AI will be able to adapt entire codebases to accommodate new features - regardless of their impact on existing technical infrastructure, but we’re not there yet..
Having an understanding of the core parts of a product’s architecture is still very helpful. And before we dig into our first practical examples together, if you’re interested in getting a little bit of background on some core architectural concepts, these Knowledge Series editions might be helpful too:
What is a microservice? https://departmentofproduct.substack.com/p/knowledge-series-13-what-is-a-microservice
What’s a monorepo? https://departmentofproduct.substack.com/p/knowledge-series-32-whats-a-monorepo
How does the release process work? https://departmentofproduct.substack.com/p/knowledge-series-29-how-does-the
Practical examples of how to use AI to understand technical architecture quickly
Now let’s take a look at some practical ways you can use AI at work to understand architecture - quickly.
We’ll start with some architectural diagrams from Meta’s WhatsApp and Snap. The WhatsApp example simpler than Snap but both contain various elements of technical architecture that are worth knowing about.
1.Translate architectural diagrams into plain English
One of the simplest ways you can use AI for technical architectural concepts is to simply ask AI tools like ChatGPT, Grok, Claude and Perplexity to describe what they see and explain it in plain English.
This diagram explains how WhatsApp preserves a user’s privacy when storing your WhatsApp contacts:
Let’s imagine you’re working on a feature relating to this and you want to get an initial understanding of this without having to book in a meeting with an engineer.
Here’s a prompt you could use: