UX for GEN AI

Evolving AI capabilities changed how I’d elevate an assistant today…

Evolving AI capabilities changed how I’d elevate an assistant today…

View solution designed in 2024

OVERVIEW

It started by reduce friction by preserving context. But how do we design memory users can trust and built over time.

I designed an AI assistant to reduce that friction by preserving context within a project.
Revisiting this work in 2026 allowed me to step back and ask a bigger question. Not just how do we help AI remember, but how do we design memory users can trust and build on over time.

2024

Reducing Repetion by Preserving Context

AI assistant that allowed users to define context once and return to conversations without re-explaining their needs. The focus was on efficiency.

2026

Supporting Continuity & Adaptability

As AI capabilities expand, users need visibility over what the system remembers. There shouldn't be freezing context at the moment of setup.

View solution designed in 2024

PROBLEM WITH ASSISTANTS

Assistants retain context, but rarely signal when it becomes outdated, making it difficult for users to trust whether responses are still relevant or accurate.

PREVENTING CONTEXT DECAY

  1. AI assistants often become outdated quietly, so the real opportunity is not automatic adaptation, but making change visible and actionable for users.

Surfacing clear signals when instructions, documents, or assumptions no longer reflect the current scope, and gently prompt users to review or update them at the right moments.

Rapid Prototype in Figma Make

TRUST THROUGH VISIBILITY

  1. Instead of hiding how responses are generated, the assistant must clearly show what information it’s using/referencing.

Providing simple, inline cues such as “Using earlier research” or “Based on draft v1” that explain where a response is coming from, helping users understand and trust the output without additional setup.

Rapid Prototype in Figma Make

THOUGHTS

Trust in AI comes from visibility. Knowing what the system is using and when it needs updating!

LEARNINGS & TAKEAWAYS

How this project is helped me grow as a designer?

01

Rapid prototyping as a communication tool

Prototyping helped me validate my thought process early by turning assumptions into tangible interactions, allowing me to reason through AI behavior and iterate with confidence.

I honed my communication skills through direct client engagement and gained the ability to understand understand and address client needs effectively.

02

Designing alongside evolving technology

Working on this project pushed me to adapt my design thinking as AI capabilities evolve. The key is focusing less on fixed solutions and more on creating systems that could remain relevant as the technology matured.

I honed my communication skills through direct client engagement and gained the ability to understand understand and address client needs effectively.

You might also like to see…

©2026 Sakshi Sonawani 🤍

©2026 Sakshi Sonawani 🤍

©2026 Sakshi Sonawani 🤍

Create a free website with Framer, the website builder loved by startups, designers and agencies.