In our blog, we explore many topics related to UX design and product management. And we thought, why not interview some cool companies, ask them how they build and manage their products, and bring you some valuable insights from the tech industry? So we reached out to Hypotenuse AI — a promising tech startup from Singapore backed by top investors including Y Combinator and other venture capitalists (VCs) and strategic angels.
Hypotenuse AI is a platform that uses artificial intelligence to write content. Machine learning helps to automatically generate content for businesses, including product descriptions, text ads, and blog posts. Their products turn keywords into surprisingly great marketing copy and recently guys implemented an AI generator of images!
Just over a year ago Lin Hui and Joshua Wong and Low Lin-Hui left their jobs at Stripe and Amazon to start Hypotenuse AI. We had a conversation with Lin and she shared her vision of product management in a young and fast-moving startup. She also gave some advice for aspiring founders.
As an agency that helps SaaS startups design their products, we at Eleken found Lin's startup management tips very insightful. We hope you enjoy this story too. Let’s not waste a minute and dive in!
The Hypotenuse AI team and roles
— Hi Lin, thanks for the opportunity to talk to you! We are very excited to ask you about the in-house processes of Hypotenuse AI. We are going to talk about product management in your startup. Though we know you are not a product manager in the classical sense, right?
— Hi, thanks for having me! That’s right, as a small but mighty team, we all wear a lot of hats and share product responsibilities, including designing the product, defining the roadmap, and managing the execution. We play to our strengths, and I work more on how the product looks and feels, while my co-founder Joshua is the go-to on architecture and technical specs.
— What are your primary responsibilities related to managing the product at Hypotenuse.ai?
— It involves quite a variety of things — collecting feedback from users, prioritizing what will be on our roadmap, from a combination of feedback and the vision that we have for the overall product.
Because we are also a growing team, there are things that aren’t in a normal product manager’s responsibilities. For example, I do product design, in collaboration with the rest of the team.
Spec-ing out features and architecture — this is where Joshua and I divide and conquer based on our strengths, as he has software and AI engineering background, whereas I have a technical sales and payments background.
Key product features and implementation
— Please, share with our readers your product vision and product value proposition.
— We’re creating a world where everyone is empowered to create content. Instead of spending your day tied up by writer’s block or agonizing over the right words, Hypotenuse AI’s vision is to allow you to create content just by describing what you want to convey and express, and not constrained by your writing skills.
The Hypotenuse AI platform was designed so that anyone, even if they aren’t an expert in writing or a particular language, can get a high-quality marketing copy automatically written for them just by describing a topic or entering a few keywords. Currently, you can use our product to generate content in over twenty languages!
— How do you implement this from a product perspective?
— We cover this in a few ways:
- Content drafts in seconds: the user inputs a few keywords, and the text is drafted in just a few seconds.
- The content is constantly improving: our AI can digest the brand guidelines, understands the customer’s brand story, and writes content in a voice that’s fluently yours. The more customers use our platform, the better it gets for them. The product learns users’ preferences and gets more personalized over time.
- Bulk management: getting and refreshing content for thousands of copy pieces like product descriptions is expensive and tedious. We integrate with product and CMS systems so that content for an entire catalog can be prepared with just a click of a button.
Agile and scrum practices at Hypotenuse AI
— How do you organize your work and management? Do you follow any Agile/Scrum practices?
— Yes, we use Agile/Scrum methodology, but we adapt it to our needs as a startup. In particular, we have one-week sprints, daily standups, and deploy new features at least once a week.
At the end of each sprint, we have a sprint planning session where we review the past week, go through specifications, assignments and estimations for each task in the upcoming sprint, and our favorite part — coming up with an outrageous name for the sprint. Last week’s name was “Ecstatic Emu”.
— Haha, I love it! And what tools do you use for product management? For example, we at Eleken rely on Figma and Notion a lot.
— Oh yeah, same here. Notion is a big one for us too in terms of organization. Our internal documentation and managing sprints and tasks also happen there.
I’d say the physical whiteboard is also one of our most useful tools. We’re an in-office team, so letting our creativity and planning run wild on a whiteboard together has been helpful for roadmapping and churning out designs and specifications.
Data plays a big part in our decisions. For this, we use tools like Amplitude, Google Analytics, and internal built systems to keep track of user behavior and feature usage.
For designing, I also frequently use Figma and low-code tools like Webflow to build our site. Starting off with a pre-built design system and UI kit is also very helpful in getting higher fidelity and polished designs out quickly, which adhere to design best practices (I think of our AI content writing tools as the copywriting form of a UI kit!)
— How big is your team?
— We’re a Lean team with just a handful of engineers. I work closely with the engineering team to spec features, though everyone is more involved than what a usual engineering role does: what the team has expressed particularly enjoying is that everyone gets to play a part in developing and creating features.
As a startup, we don’t really break into departments beyond the tech and business teams (folks tend to multi-hat here).
UX Design and product management
— That sounds amazing! As a team of product designers, we are curious to know how you guys make UX design decisions in your team.
— I typically design most of our frontend, but features are discussed with the person working on the task first. As we move quickly, not everything is always designed to the highest fidelity as well.
There is space and opportunity for every member of the team to express and think through features more holistically, ask good questions and build with the user in mind, rather than just execute to the exact specs provided.
— Totally agree here, we also like the holistic approach to product design. What about UX research?
— Everyone on the team does user research to different degrees, from observing user sessions and reaching out proactively for feedback. We have a central document for creating user feedback. It can get a little unstructured at times, but it serves as a quick and central location for the exact words from the user.
— How do you usually organize work on new features?
— There are typically two sources of new features for us: the roadmap to our vision, and what users ask for (whether implicitly or explicitly). Ideally, both of these should overlap heavily. We decide what goes into each sprint by balancing between user impact, urgency, and resources required (mainly engineering time).
After we've decided to work on the feature, we’ll break it up into different tasks. We add a priority level and estimated the size of the task using Fibonacci numbers before assigning an engineer who will be working on it for the next sprint.
Depending on the feature being worked on, we may do a full Figma design for it (especially if it’s frontend heavy), and/or a technical design document that includes the business logic, system architecture design, and interactions with other components.
— What are the specifics of building a startup in terms of design and product management?
— We take a lightweight approach to design and planning to balance speed and scalability. An ideal design document is a multi-page and neat document that considers all the SWOT risks (Strengths, Weaknesses, Opportunities and Threats analysis), data analysis, and every detail down to the pixel. But we often find diminishing returns on full granularity when we’re iterating quickly in a startup.
For example, for frontend tasks, full Figma designs are only done when UI changes are significant. Instead, most of our backend heavy architecture designs are just a deeply focused one-hour conversation on a whiteboard, which we photograph and save in our documentation. This ensures that key considerations are still thought of, while still allowing us to move fast.
Lin’s favorite resources for inspiration
— Lin, what do you read? Please share the books and resources that help you in your work.
— For writing, I like reading David Perell and Lawrence Yeo from More To That.
For the product, Julian Sharpio has clear, concise tips, and I really enjoy the animated case studies from Growth.Design. I also use Matter to discover long-form content — it’s one of my favourite apps!
— And our last question is what would you recommend to aspiring founders? What to focus on, where to start, any advice.
— There are a lot of things to consider when starting a business, and it largely depends on what you want to achieve with it. Aside from the journey of improving your product-market fit constantly — if you want to go big and venture scale, then you need to focus on building a strong team and scaling quickly. But if you want to bootstrap and run a smaller business, then you need to think about things like how you're going to grow sustainably from the beginning.
No matter what you choose, there are a couple of things I’ve personally found useful to focus on:
- Find something to be curious about in everything. There is a lot, past and present that we can draw from the world around us and apply to very different subject matter areas. Being curious also flattens your learning curve on any topic, because you just want to learn more about something for the fun of it, not just to go through the motions.
- Related to curiosity is being a “focused generalist”. When you’re curious enough about things to get past the initial valley of difficulty, just being good enough to be dangerous in 2-3 areas is already an advantage, rather than being a jack of all trades. Find a project to work on and practice.
- This is already often said, but launch your product as early as you can. This will give you the opportunity to get feedback from actual customers and make changes based on that feedback.
— Lin, thank you so much for the great talk. We wish Hypotenuse AI a bright future. And if you ever need skilled and experienced designers to develop your product further, we will be happy to help! Just contact us.
— Haha, alright! Thank you, guys. Good luck to Eleken team and to all the readers.
