Design Lead Process Ongoing

Integrating AI into how we design.

My team had been working AI into our design process, and my own results stood out enough that Havells asked me to put the thinking into a presentation they could share with their design head.

Role
Design Lead
Context
Internal workflow
Status
Ongoing

Taking stock of how we were using AI.

AI was already part of how the team worked. What did not yet exist was a clear picture of where it was genuinely helping, how much, and what fuller adoption would look like. This work was about taking stock of how we were using AI across the design process, measuring the difference it made, and setting out a plan to build on it.

Built into each stage, and measured.

Rather than treat AI as a separate tool, we worked it into the existing process stage by stage. These are the measured results from our own workflow.

Stage
Before
After
Discovery & research
Manual competitive analysis, 2 to 3 days
Synthesis in 4 to 6 hours
Scoping & PRDs
Half a day to draft
Around 2 hours, edge cases caught early
Ideation
One direction at a time
Multiple directions in minutes
Design production
Full screen from scratch, hours each
AI-first drafts, designer refines

Standout: a full UX audit and redesign of the PanSol Admin portal, covering audit, PRD, UI, and prototype, done in about 2 days against roughly 2 weeks of manual work.

The next phase.

These are projected, not yet implemented.

Spec writing From a full day documenting states and dev intent, down to around 2 hours.
Design-to-code A shared component language between design and engineering, with fewer review rounds.
Capacity With fuller adoption, room for roughly 2x more projects in a year.

Faster, and noticed outside the team.

  • PanSol audit and redesign: 2 weeks of manual work down to about 2 days.
  • More time on the work that matters: strategy and research, less on production.
  • An external signal: after seeing how the team worked, Havells asked me to put together a presentation on our AI use and adoption strategy, to share with their design head and use internally.

Ongoing.

This is an active effort, not a finished project. The measured changes are in daily use. The part still open, and the most interesting, is working out how to collaborate with developers using AI, so design and engineering reach the most efficient end result together rather than each using AI on their own.

More work