Sarvam.ai
Designing AI Experiences for the
world’s most diverse market
Sarvam is building foundational models trained in Indian languages, for Indian use cases, making generative AI radically more inclusive. Backed by Lightspeed, Peak XV, Microsoft, and anchored in the public digital infrastructure movement supported by Nandan Nilekani, they are now also partners with the Government of India to develop its sovereign LLM, actively shaping the backbone of India’s AI future.
Industry
Artificial Intelligence
Stage
Early-stage
Deliverables
AI Experience Design, User Interface Design,
Usability Testing
The vision
Sarvam wasn’t just building models. They were building the future infrastructure for how 1.4 billion Indians would interact with AI in local languages, with cultural context, and real-life constraints. From household shopping to health dashboards, their ambition was to make AI not just accessible, but usable and trusted by everyday users across India.
The challenge
Sarvam was still in its early product-market fit phase, with powerful models but no live interface to anchor them. They needed help figuring out where to apply their capabilities, how to surface them across products, and what the experience should feel like.
There were no templates to follow, this was zero-to-one work for a category that didn’t exist anywhere else in the world.
The solution
We partnered with Sarvam during their exploratory phase to shape what Indic-first AI products could look and feel like. Over a quarter-long engagement, we ran structured sprints across e-commerce, finance, healthcare, and entertainment to identify high-potential use cases and prototype user-facing experiences grounded in trust, utility, and cultural context.
From ‘Models’ to
‘Moments’
We developed a multimodal design framework that helped the Sarvam team evaluate the right way to surface AI in different contexts - inline assistants, floating overlays, proactive companions, full-screen immersive views.
Each format was paired with guidelines for when to automate, when to assist, and how much control to hand back to the user.
Switching modes should be easy and intuitive
Help users understand what all the system can do.
Design for flexibility in unpredictable moments
Financial advice that sounds like a friend
From savings goals to impulsive purchases, we prototyped a conversational “budget buddy” that could help users plan better, spend smarter, and recover from financial setbacks without judgment. We also designed contextual prompts to adjust investment plans in real-time and nudges that respected both user intent and lifestyle trade-offs.
Healthcare that speaks your language
We reimagined healthcare with a focus on clarity, personalization, and reassurance. Our prototypes included AI-assisted medical summaries, conversational “what happens if…” explainers, medication history visualizations, and simplified clinical documentation - all designed to information more accessible to both patients and providers.
Booking experiences that predict intent
We reimagined the travel booking experience to feel more conversational, predictive, and human. With flows where AI models could infer traveler intent through natural language inputs - from pre-filling destinations and dates, to surfacing flight options dynamically as preferences are shared.
Designing trust, not just interaction
One of our core contributions was a decision-making framework that helped Sarvam product and design teams evaluate when to use proactive AI, how to balance automation vs. control, and how to handle edge cases like fallback, error, and escalation.
Designing a future
before it’s here
This wasn’t a traditional delivery engagement. Nothing shipped and that was the point. Our work helped Sarvam:
- Prototype AI experiences across high-priority verticals
- Align internal teams on what good AI UX looks like
- Establish a design language that could scale with the tech
- Move beyond demos into user scenarios that were real, valuable, and Indic-first
Reimagining what
AI feels like
At Parallel, we believe technology is only as powerful as the experience it enables. With Sarvam, we weren’t designing UI, we were designing trust. Across industries, languages, and contexts, we helped shape what it means for AI to belong to a user. Not just technically. Culturally.
