Circle to Search - Part 3

Circle to Search, a vision still shaping the road ahead

Circle to Search began as a new way to search anything on your screen. This vision looked at what could come next: helping people get a clear answer, work through a problem step by step, and act on what they find. It started with research, and some of this thinking went on to guide what's in the product today.

Vision · Design sprints · Contextual AI

2

Design sprints

3

Vision pillars

1

Shared roadmap

Context

The vision started with what people actually needed

After the first launch, I led a series of design sprints to explore, prototype, and test new ways to meet the needs our research had surfaced. The goal was a clear vision for the near future: a high-fidelity story, grounded in research, brought to life with working demos.

The research

The vision started with what people actually needed

Before designing anything, I looked at what people were really asking Circle to Search for. The needs sorted into clear tiers, from the biggest unmet opportunities to things the product already did well. This is what pointed the vision toward where it could help most.lp most.

Biggest oppurtunity

Get summary or key points
Troubleshoot tech issues
Pros and cons list
Ask AI follow-up questions
Return to past searches
Ask contextual questions
Verify scams

Room to improve

Verify AI presence
Create custom image
Share search results
Save search results
Homework help
Identify famous people

Lower priority

Reminders to use it
Listen to screen content
Annotate content
Video game help
Mobile game help

Already well served

Scan QR or barcode

Save part of screen
Share part of screen
Copy and paste a

From needs to pillars

Three things people come to do

Underneath all those needs were three real intents. Some people want a clear answer. Some want to be walked through a problem. Some want to act on what they found. I built the vision around these three.

Pillar 1

Know

User Goal: Get a clear answer to what's in front of me

Example: Land on an unfamiliar article and want the gist before reading

Needs:

Get the key points
Weigh pros and cons
Check if something's trustworthy
Identify what's on screen

Pillar 2

Work it out

User Goal: Get walked through something I can't know at a glance

Example: Stuck on a tricky problem and want to understand it, not just see the answer

Needs:

Homework help
Troubleshoot a tech issue
Work through a problem
Step by step
Ask follow-up questions as you go

Pillar 3

Act on it

User Goal: Take the next step from what they found

Example: Found the thing, now want to share it with friends or come back to it later

Needs:

Save results
Share results
Create an image
Return to past searches

Pillar one

Know: get a clear answer

Sometimes the answer is all you came for. This part of the vision handed it over cleanly: a summary of a long page, the key points, a quick check on whether something can be trusted. You ask, you get what you need, you move on.

Long press home & talk

Results appear

Long press home

Results appear

Pillar two

Work it out: get walked through it

Some things don't have a single answer to hand over. They have to be worked through. So this part guided people step by step, through a tough problem, a tech issue, a question that needed a follow-up. The point wasn't just to solve it, but to leave you understanding it.

Rich homework help

User draws

Long press home

Results appear

Pillar three

Act on it: take the next step

Often the answer isn't the end. It's the start of something. This part helped people do the next thing: save it, share it, act on a number or address, pick a search back up later. The result became a step, not a stop.

Long press home & talk

Contextual results appear

Take actions

Pick up where you left off

Designed to fit the moment

Help that matches what you came to do

Open it over a game and you probably want a tip, not a product. Over a recipe in another language, you want to read it. The vision tuned the help to the moment: game tips on a game screen, live translation when there's another language, the right action for what's in view.

Results appear

Long press home

Results appear

Game tips

Better discovery of capabiities

Open it over a game and you probably want a tip, not a product. Over a recipe in another language, you want to read it. The vision tuned the help to the moment: game tips on a game screen, live translation when there is another language, the right action for what is in view.

Introducing new capabilities

Introducing new capabilities

From vision to product

Some of this is already in the product

The helpful suggestions above, one of the first ideas from these sprints, shipped to people. And as AI keeps moving forward, more of what we imagined here becomes possible.