Product thinking, practical voice workflows, experiments, and the things we’re learning while building Lungoor.
No content treadmill. Just things we learned, tested, or changed.
We hesitate, switch languages, ramble, correct ourselves and change direction mid-sentence. Good voice software should understand that behaviour instead of training it out of us.
Voice at WorkKshitija Katare6 min readRead the thinkingShort reads on voice-first work, product decisions, and the systems behind a calmer way to work.
People pause, restart, mix languages and change their minds halfway through a sentence. Voice products should not treat that as bad input. That is simply how thinking sounds.
Kshitija Katare6 min readBuild NotesRemoving “umm” is easy to demonstrate. Preserving intent while cleaning speech is a much harder engineering problem.
Chetan Divekar6 min readBuild NotesA tiny widget that appears anywhere on your desktop sounds simple. Making it feel reliable across apps, states and workflows is anything but.
Haider Patanwala7 min readInside LungoorThe goal is not to build the loudest AI product in the room. It is to build something useful enough that people notice when it is missing.
Nishant Patel5 min readBuild NotesThe Lungoor widget has one difficult job: show enough to create trust, then get out of the way before it becomes another surface to manage.
Yuvraj Rane4 min readInside LungoorBig launch moments are visible. Product quality is usually built in the much quieter rhythm of feedback, fixes, decisions and follow-through.
Monil Shah4 min readNew product notes, useful voice workflows, and things we’re learning. No daily inbox occupation.
Useful ideas. No spam. Ever.