Building Generative AI Products Case Study: Uizard

Welcome to the latest issue of the Product Management Learning Series - a series of live streaming events and newsletter articles to help you level up your product career! ??

I’m launching a special series sharing what I learned from various companies building the latest and brightest generative AI products. Today, please welcome our guest, Tony Beltramelli , the co-founder of Uizard by Miro Labs , the groundbreaking AI-powered UI design tool that is reshaping the landscape of digital design. In a landscape dominated by giants like Figma and Adobe, Uizard has carved out a space of its own, offering rapid, intuitive design solutions that transform ideas into prototypes in mere minutes. Dive in!?

If you missed our conversation, you can watch the full recording here. ?

Shyvee: Can you share the story of how the idea of Uizard came to be? What was the initial problem that you set out to solve?

AI summary of Tony's response: Tony started Uizard over 5 years ago in early 2018; he and his co-founders were struck by the inefficiency in the traditional software design and development process. The cycle of converting ideas to pixels and then to code was entirely manual, a process they found archaic and ripe for disruption. Recognizing this gap, they envisioned a better way to streamline this process through AI, long before Generative AI became a buzzword. The aim was to make it easier for people to experiment with designs, create interactive prototypes to test ideas quickly, and ultimately make the entire product development process more efficient and intuitive.

In the live demo, Tony showcased Uizard's powerful AI-driven features designed to accelerate the design process for product teams. Uizard's "Autodesigner" allows users to generate projects without starting from scratch, making it easier for those who are not trained designers but understand customer needs. Another unique feature Tony highlighted was Uizard's ability to convert hand-drawn wireframes or even screenshots into editable, interactive digital screens using computer vision and AI. This enables teams to iterate quickly on existing products or prototypes. The platform's drag-and-drop interface makes the entire process user-friendly, allowing for real-time edits and interactions. Watch the full demo here and check out Uizard yourself!

Shyvee: Walk us through the product development journey. How did you go from idea to a functioning product?

AI summary of Tony's response: Tony outlined the unique development journey of Uizard, emphasizing that unlike other startups, they didn't start with user interviews. Instead, they focused on building a "minimal viable AI," as the core idea was to leverage AI to solve design problems that traditional tools like Adobe and Figma couldn't address. Once the AI was in place, they launched and began collecting user feedback. Tony admitted that the product had its shortcomings during the alpha phase, which lasted two quarters. It took about four years from inception to launch, a timeline he noted was similar to Figma's. Throughout this period, the Uizard team underwent numerous iterations, continuously learning from customer feedback to improve the product. They maintain a rapid shipping schedule to this day, always seeking to better fit their customer's workflows.

Shyvee: How did you ensure that Uizard remains user-friendly and amplifies human creativity while incorporating advanced AI functionalities?

AI summary of Tony's response: Tony emphasized that Uizard is not designed to replace human designers but to facilitate better communication between product managers, developers, and design teams. He addressed several critical design principles to maintain user-friendliness while integrating advanced AI. First, Uizard allows users to modify AI-generated outputs. Tony notes that AI will inevitably make mistakes, and defining user intent is non-trivial; therefore, human-in-the-loop involvement is crucial. Second, the tool supports real-time collaboration, acknowledging that product development is a team effort. Third, Uizard abstracts design complexities by allowing users to work with higher-level concepts instead of just pixels and vectors. This makes the tool accessible to non-designers who can think in terms of buttons, input fields, and sign-up flows. Lastly, Tony mentioned that accessibility is paramount to ensure the tool is easy for everyone to use.

Shyvee: Can you share some insights into who your early adopters were and how you reached them?

AI summary of Tony’s response: Tony revealed that Uizard initially attracted entrepreneurs and early-stage startups as its first adopters. However, over the past 18 months, the tool has gained traction among product leaders and teams in large, well-known organizations. The primary growth channel for Uizard has consistently been word of mouth. The company focused on leveraging this organic growth by building engines to accelerate it, such as incentivizing invites and shares through a reward program. They also employed community-led growth strategies, encouraging users to share their experiences on social media. While they've tried to influence these facets of word-of-mouth, Tony emphasized that the ultimate driver remains the product's quality. The goal is to make Uizard so compelling that users will want to share it without any external prompts.

Shyvee: How did early user feedback shape the evolution of Uizard? What were the surprising insights?

AI summary of Tony’s response: Tony shared that one of the most surprising aspects of early user feedback was the speed at which users integrated AI into their workflows. Despite initial skepticism about AI's limitations, Uizard found that generative AI shortcomings such as hallucinations were actually features, not bugs, that spurred innovation and enabled new ideas to emerge. However, not all was smooth sailing. Tony mentioned that some features had to be killed off because they weren't resonating with users. For instance, they had built AI models to generate code, thinking it would bridge the gap between design and development. Despite spending significant time on this, they found that users weren't interested. Companies had their own engineering guidelines and code standards, making the feature redundant. This led to the painful decision to remove these features, underscoring the lesson that customer feedback is invaluable and, ultimately, the customer is always right.

Shyvee: Let’s talk about Uizard’s GTM strategy. Can you discuss the thought process behind Uizard’s pricing strategy?

AI summary of Tony’s response: Tony delved into Uizard's pricing strategy, revealing that the tool was initially free to encourage user adoption and gather feedback. Pricing was introduced in February 2021 when the product was launched out of private beta. Pricing is structured to cater to different user groups. A free plan aims to attract students and early-stage startups, acting as a funnel to convert them into paying customers later. The Pro Plan is designed for established startups and product teams who require more AI usage and premium templates. For larger enterprises that require heavy use of the design tool, a Business tier is offered.

Initially, Tony and his team dived into data analysis and customer surveys, including trying the famous Van Westendorp price sensitivity method, only to find that the most intuitive decision—pricing alongside competitors—was the correct one. Tony advised product builders to trust their intuition and be prepared to iterate on pricing over time. To learn more, read The Ultimate Guide to Pricing your SaaS Startup with Data directly from Tony!

Shyvee: How does Uizard position itself in the competitive landscape with giants like Adobe, Figma and Canva in the market?

AI summary of Tony’s response: Tony addressed Uizard's position in the competitive landscape by emphasizing its specialized focus on AI for design. Unlike giants like Adobe and Figma, which excel at serving designers, Uizard aims to fill a gap by catering to product managers and business stakeholders who understand business needs but may not have design expertise. He mentions that Canva has done well by incorporating generale-purpose AI, but Uizard's advantage lies in its vertical-specific, best-in-class AI tailored for design. Built from the ground up, this specialized tech offers a first-mover advantage that becomes stronger as more customers and usage data come in. This makes it increasingly difficult for competitors to duplicate their models. While Uizard does integrate with other AI models to offer the best value, its core strength remains its specialized focus on empowering non-designers to collaborate more efficiently with design teams.

Shyvee: What are some “moats” Uizard are building to defend itself for long-term sustainability?

AI summary of Tony’s response: Tony outlined multiple "moats" to ensure Uizard's long-term sustainability. Technically, the most significant moat is the specialization of their AI stack for design. This is fortified by unique data layers, proprietary models architected and engineered by their team, and the flexibility to integrate third-party models. Uizard has a unique advantage in having a AI specialized in UI/UX design, making it difficult for competitors to replicate unless they invest in intensive machine learning research.

From a distribution standpoint, word-of-mouth has proven to be a strong channel for them. Tony acknowledges that while some debate the efficacy of word-of-mouth as a sustainable moat, once a product is compelling enough to talk about, it creates a lasting impact. Despite these moats, Tony emphasized the need for vigilance and even a degree of paranoia, acknowledging that moats can always be crossed with the advent of new technologies. Therefore, maintaining a best-in-class offering is crucial.

Shyvee: How do you envision the future of design with AI, and what role do you hope to see Uizard play in that future?

AI summary of Tony’s response: Tony highlighted the rapidly evolving landscape of generative AI, urging the need for continuous innovation to maintain a best-in-class product. He sees a transformative change on the horizon for creativity, impacting various forms of content from audio and video to design. Tony is particularly excited about the future of design maturing into specialized domains, akin to how software engineering has evolved into various focused disciplines like front-end, back-end, mobile development, etc.

He believes that AI will catalyze this specialization by enabling domain-specific experts to incorporate design into their workflow. Uizard aims to be a part of this transformation by helping product managers and builders become specialized designers within their domain of expertise. The goal is to leverage AI to serve a more diverse and specialized audience in the design realm.

Shyvee: What are some of your personal reflections building Uizard??

AI summary of Tony’s response: Tony candidly acknowledged the volatile and fast-paced nature of the tech landscape, admitting that it's a rollercoaster of emotions and "freak-out" moments when new technologies emerge. He shared several significant learnings from building Uizard. First, he stressed that things always take longer than anticipated. Drawing a parallel with Figma, he highlighted that what appears to be an "overnight success" usually involves years of hard work.?

Second, he advised "killing your darlings"; if a feature doesn't add value, it should be removed promptly. On the topic of pricing, Tony expressed regret over delaying the introduction of a pricing model. He noted that paying customers often provides better feedback. Finally, Tony pointed out that traditional startup advice doesn't always apply to AI startups. For instance, the common Y Combinator advice to launch even if "you're embarrassed" by your first version doesn't hold in the AI space. In AI, the core technology must be robust enough to draw initial user value, deviating from the usual lean startup methodologies. The interface through which users engage with the AI core may exhibit initial imperfections at launch, but it is imperative that the core AI is a robust foundation to provide genuine value to early adopters.

I hope you enjoyed reading my conversation with Tony and learned something from it. I started this special series as an opportunity to learn more about smart people building interesting products in the AI space. I’m excited to learn more about their process and the technology/tools they’re using, the big problems they are solving to make an impact on people’s lives, and the hard questions they are asking themselves to grow sustainably.??

If you have any questions, feedback, or suggestions on how to make this series better and interesting people or companies that I should talk to, please reach out! I’d love to hear from you.??

Thanks for reading,?

Shyvee

Next up, join us for an exciting conversation with Vivek Patel , Yelp's former CPO and an influential product leader for over a decade. Vivek navigated Yelp’s exponential rise from a startup generating $20M in revenue to a $1B+ industry leader, serving over 33M mobile users and 500K advertising locations. Now an advisor, Vivek offers invaluable insights for those looking to excel in product building, AI, and career growth.

?? RSVP here (Replay available upon registration)

Next up,

• Nov 3: A Masterclass in Product & Career Scaling with Vivek Patel, ex-CPO of Yelp
• Nov 10: Revolutionizing Creation on Roblox with Generative AI
• Nov 17: A Master Class in Navigating Enterprise LLM Challenges with Mahesh Yadav
• Dec 15: Mastering the Art of Decision-Making & Execution with P.J., Thumbtack CPO

More sessions coming… subscribe to stay tuned!?

Learn more about the Product Management Learning Series and view past recordings here.