Lex: AI Writing Assistant: July 26, 2026

Lex AI Writing Assistant

Participants

  • Sandeep Garg – Bridgetree, Noida | Deliverable
  • Maneesh Maddala – Morningstar Inc, Chicago | Deliverable
  • Manish (Manish Kumar G) – Thomson Reuters, Bangalore | Deliverable
  • Dhrumil Soni – Tech Holding, Ahmedabad | Deliverable
  • Bharath S – Rocket India, Chennai
  • Vishnu Priya M – Andinolabs, Chennai/Bangalore | Deliverable
  • Ganesh Manoharan – TransUnion (Sr QA Engineer) | Deliverable | Deliverable
  • Rudreshwar Jha – KritiKal Solution Pvt. Ltd.
  • Maniksha Bobby – TCS, Bangalore | Deliverable
  • Shrinathacharya L M
  • Dhanu Arumugam – TransUnion, Chennai
  • Trisha Chetani | Deliverable
  • Dharmendra Kumar | Deliverable
  • Pawan
  • Ajay Balamurugadas – Facilitator

Charter

Explore Lex.page’s core areas using a fresh account, to model key features and produce a short testable feature list.
Timebox: 40 minutes, followed by a group discussion.

Date

July 26, 2026, 6 PM – 7.30 PM IST

Video:

Link

Brief summary of interesting findings/views

Several testers converged on similar friction points: multiple participants (Maneesh Maddala, Ganesh Manoharan, Dhrumil Soni) said the biggest challenge was covering all of Lex’s features within the 40-minute timebox, often at the cost of deeper exploration. Ganesh reported a functional bug — the app appeared to hang or stumble upon reaching a maximum word limit, though it self-recovered within about a minute — and separately mentioned that the free tier runs on a model called “gpt-5-mini.” Maneesh noted specific UI friction around the archive/unarchive menu and keyboard shortcuts, which he wasn’t able to fully test in time.

Did you find the bug on this page 🙂 Yes, the button – Download Transcript is broken! (WIP)

On the reflective side, Sandeep Garg offered a more conceptual take, framing Lex as a purpose-built “wrapper” around GenAI rather than a general-purpose assistant, and expressed interest in further testing its factual accuracy and generation/correction workflows. Several others (Manish, Trisha, Dhrumil) described initial nervousness or time pressure but ended the session feeling reasonably positive, with common takeaways being a desire to test more advanced or “agentic” features in a follow-up session. Overall, the tone across responses was constructive and exploratory rather than critical — most treated it as a learning exercise in time-boxed exploratory testing rather than a rigorous product evaluation.