Nishtha

VP of Design & Research @Workato. Building products and teams. Making complex workflows and AI agents easier to use.

[Nishtha @nishtham4]

@panthshah_ @krispuckett I’m glad you enjoyed it!

[Nishtha @nishtham4]

A week ago, when @krispuckett offered to review portfolios, he found my website’s hero a bit buggy. That gave me the perfect excuse to try something new.
Made a little game about pesky agents breaking boundaries.

[Nishtha @nishtham4]

@Mo_webflow @krispuckett Haha, not a bad outcome if they find it engaging!

[Nishtha @nishtham4]

@elenahuxy @krispuckett It was equally fun making it!

[Nishtha @nishtham4]

Portfolio link: https://t.co/X7tyJFBlUD

[Nishtha @nishtham4]

@isablyns @krispuckett Thanks!

[Nishtha @nishtham4]

I used Codex to help build it with Next.js, React, TypeScript, and Canvas. The movement uses spring and collision simulations, so the boundaries stretch and wobble, characters jump and fall with gravity, and drawings become obstacles. I used Qwen to interpret what people draw and pick a response from a set of behaviors I’ve defined.

A lot of the design work was in making the reactions feel playful. It still feels pretty rough to me, but I wanted to put it out in the world rather than iterate on it forever.

[Nishtha @nishtham4]

@krispuckett I had so much fun making it! Given that my son got obsessed with drawing over it, I might turn it into a proper game for him to play 😃

[Nishtha @nishtham4]

@krispuckett Thank you for taking the time, Kris! So helpful to see it through another designer’s eyes. Pushed some updates based on your comments.

[Nishtha @nishtham4]

@tikhbana So true! I’ve been trying to find good examples of in-depth product thinking led by design, but my feed is full of shiny micro interactions

[Nishtha @nishtham4]

@seb__design Some of the best designed things are work done by super talented people who don’t have bandwidth to be on X. I joined X recently and I’m blown away by the half hearted dizzying work that gets kudos here.

[Nishtha @nishtham4]

@PranavOriginals The footer explorations you have are interesting!

[Nishtha @nishtham4]

Level 5 of my AI maturity model is Autonomous: AI would research, design, evaluate, and prepare a spec. We write the brief and approve the result. We become the client; AI becomes the design agency.
Full framework: https://t.co/6OUssvMgjv

[Nishtha @nishtham4]

Level 4 of my AI maturity model is Intentional: AI helps create emotional resonance. We guide it with taste, context, and critique. The irony is that as soon as AI reaches what feels right, our expectations shift. The ceiling keeps moving.
Full framework: https://t.co/6OUssvMgjv

[Nishtha @nishtham4]

Once we have options, we need to question them. Level 3 of my AI maturity model is Evaluated: AI looks for patterns in the evidence and challenges our thinking. We set the rubric, weigh the evidence, and decide what changes.
Full framework: https://t.co/6OUssvMgjv

[Nishtha @nishtham4]

Next, we trust AI to explore a problem. Level 2 of my AI maturity model is Generative: AI becomes a brainstorming partner, proposing different directions. We choose what’s worth pursuing and cut the rest.
Full framework: https://t.co/6OUssvMgjv

[Nishtha @nishtham4]

In my last post, I asked what we’re trusting AI to do. Level 1 of my AI maturity model is Mechanical: we set the task, provide context, and review every output. AI executes. The task gets faster; the workflow stays mostly the same.
Full framework: https://t.co/6OUssvMgjv

[Nishtha @nishtham4]

@clairevo @chatprd @googledocs Nice! Each critique tied to the exact passage makes it much easier to discuss the decisions in context.