Hi, I'm Nishtha.

I lead design and research at Workato, where I'm building the platform that gives a flock of unpredictable AI agents their rules and shape.

Hands-on design

Designing Airo — then outgrowing it

We shipped a new way to work with AI in 2025 — then multi-agent architecture made it obsolete.

Co-inventor

Can streamed AI responses have conversational etiquette?

We replaced a jittery, ambiguous response stream with a patented conversational cue that shows when AI is still responding and when it is the user's turn.

Design direction

The peek-a-boo agent bar

Running several AI agents while you keep doing something else needs a UI that stays out of the way until it doesn't. We tried three placements before finding the one that fits at the bottom of the screen.

Team leadership

Agent Studio: two products, one model

Consolidating two workspaces into one platform, guided by research findings that spared no one.

Design direction

Canvas: the picture on the box

Enterprises had the pieces of their automation but not the whole picture. Canvas is that picture — and it did not arrive the way anyone first pitched it.

Executive sponsorship

MCP apps: shipping outside the roadmap

A small team with permission to skip the process, shipping MCP capabilities ahead of any roadmap.

Design direction

Retooling a design org for the AI era

I didn't tell 22 designers to "use AI." I rebuilt what they could build, and what they believed possible.

Hands-on design

Giving an enterprise product a face

A custom illustration system built to make Workato feel designed for humans, not just enterprises.

Some tinkering. A few musings.

The mirror

The murmuration

Recipe editor

Capability to product

AI Maturity Levels

Agents, not features

Voice, trust, machines

The disposable prototype

Nishtha presenting on stage at a Workato conference

I've been leading Product Design and Research at Workato for the last 10 years. In that time, design has scaled from one person to a 22-person global team. I'm quietly proud that it's a team almost nobody leaves.

Along the way, I've designed the same platform through three eras: integration workflows, copilots, and now agents that act on their own.

I'm formally trained in physical product design, visual communication, psychology and human factors. I've been researching how people learn to trust machines since before that question had an industry attached to it.

I'm looking to build with companies, in software or hardware, where the hard part is making powerful systems people can trust.

LinkedIn comment from Vijay Tella, Co-founder and CEO of Workato: “Nishtha Mehrotra - I'm incredibly grateful—we're so lucky to have you leading the intuitive product experience we're known for. I've spent 30 years in integration middleware, and this tech is complex and whack under the hood. Refactoring six complex platform categories into one while making it seamless for both pro-code integration experts and business teams? No joke. What you and your team have built is one of the great feats of enterprise software design. So much so - you have created a marketing challenge as Workato confounds people. How can something that looks so easy be so powerful? What blows me away is your first-principles thinking and relentless drive to keep attacking the complexity. We don't have real competition for our product experience, productivity, and speed—our biggest challenger is ourselves. You and your team are awesome role models of our core value to keep raising our game. Thank you.”

From Vijay Tella, Workato’s CEO

Can streamed AI responses have conversational etiquette?

Role
Co-inventor
Team
Nikhil Venkatesh
Timeframe
2023
Outcome
Design patent granted (US D1,108,461 S)

Streaming has no turn-taking cues

We launched Workato Copilot in 2023. It generated structured responses inside our recipe and connector builders, often combining instructions, outlines, and blocks of code.

These responses arrived as a stream, so the chat bubble kept changing shape as the model produced more text. The interface felt unstable and jittery. When generation paused, the bubble also appeared complete, even when more content was still coming.

In human conversation, tone, pauses, and body language tell us whether someone has finished speaking. We wanted to bring those same cues into the chat interface.

The interaction needed to satisfy three conditions:

  • Make it clear that the response is still forming.
  • Make completion noticeable without demanding attention.
  • Make the transition feel conversational, not like system status.

The proposal

While a response is streaming, its left edge is defined while its right edge fades into the background. The bubble looks unfinished because it is unfinished.

When generation ends, the border resolves around the complete message. After a brief pause, color fills the bubble and the input becomes active again. Together, these cues signal that the model has finished speaking and it is the user's turn to respond.

Before and after

Hi! As your copilot, I'm here to help you build your connector. To start, please provide the application's API documentation URL — this helps me build code for this connector better and faster.
https://stripe.com/docs/api/usage_records/create
Let's first add a connection! This allows the connector access to the API to test actions and triggers.
Add a connection
What authentication do you want this connection to use?
Private access token
Ask a question about your code
Hi! As your copilot, I'm here to help you build your connector. To start, please provide the application's API documentation URL — this helps me build code for this connector better and faster.
https://stripe.com/docs/api/usage_records/create
Let's first add a connection! This allows the connector access to the API to test actions and triggers.
Add a connection
What authentication do you want this connection to use?
Private access token
Analyzing…

Top: Before. Bottom: After. The same conversation, with the same delay in streaming. Above, a conventional bubble appears finished whenever the stream pauses. Below, the unfinished edge makes it clear that the response is still being written.

Why this detail mattered

It was a small change, but it improved the rhythm of the interaction. People could tell whether Copilot was still working without reading a status label or interpreting an indefinite spinner.

The pattern shipped in Workato Copilot's recipe and connector builders.

The underlying system was filed as a U.S. utility application (2025/0047627 A1), and the transition itself was granted as a U.S. design patent: Display Screen or Portion Thereof with a Transitional Graphical User Interface (US D1,108,461 S).

The granted design patent — U.S. Patent No. D1,108,461 S, a continuation of the original utility application, covering the transition itself as an ornamental design.
The granted design patent — U.S. Patent No. D1,108,461 S, a continuation of the original utility application, covering the transition itself as an ornamental design.