Sridhar DhulipalaFounder, SAX · Principal Advisor (Design), UIDAI

I design how people, AI agents and institutions work together.

  • 1997India’s first internet bank. I designed it at Infosys.
  • 200M+Learners on DIKSHA, which grew from the EkStep platform my lab designed.
  • 69.7MDownloads of the new Aadhaar app, Jun 2025 – Oct 2026. I advise UIDAI on its design.
Book 20 minutes
One idea, two faces

Execution is cheap now. Understanding is scarce.

Two kinds of understanding are scarce. A citizen needs to understand what a public system is asking her to agree to. An enterprise needs its people to know its business, now that AI agents write much of the code. I design for both, and I build what I design.

  • For enterprises · context

    The New Enterprise

    Clients now question the person-month. I help IT services firms, capability centres and their clients price, staff and measure work by how well their people know the client’s business. AI agents do the codified work, under signed mandates.

  • For public systems · comprehension

    Kya likha hai?

    “What does this say?” India records a person’s consent. It does not record whether the person understood. I design public systems that check understanding before a consequential act, in the person’s language and by voice.

Manifesto · 16 October 2026

Execution is cheap.Context is dear.Every action can be proved.

Read The New Enterprise
Selected work

I see the grand picture, and I give it legs.

Each case pairs the idea with what shipped.

  • 2024 – now · UIDAI

    The new Aadhaar app

    The grand picture

    Billion users, million journeys: one canonical journey for every resident, told a million ways, each fitted to the person in front of the screen.

    The legs
    • Design behind the new Aadhaar app, as Principal Advisor (Design) to UIDAI
    • Directed the redesign of uidai.gov.in
    • Selective disclosure: a resident shares only the field a transaction needs

    69.7Mdownloads and 13.8M updates by users, 21 Jun 2025 – 14 Oct 2026

  • 2026

    An AI-first national DPI

    The grand picture

    A national public system designed AI-first: inclusive for people who prefer their own language and voice, efficient to run, and owned by the agency that runs it.

    The legs
    • Architecture of QR-code credentials that can be verified offline, building on my UIDAI and EkStep work
    • Verifiable, guardrailed AI content generation with adversarial quality checks, through a technology start-up I brought into the architecture
    • A design system and a multilingual voice assistant
    • Built through AI coding agents, then handed to the owning agency with its architecture, deployment and next steps documented
  • 2021 – 2026

    The New Enterprise

    The grand picture

    By 2031 leading enterprises will run on three facts: execution is cheap, context is dear, and every action can be proved. Agents carry the proof; people carry the context.

    The legs
    • A manifesto: twelve theses, five contracts and eleven measures
    • An A1 poster of the architecture, as a service blueprint with five lines
    • An essay, The New Arrangement of Collaboration, on how IT services firms can price context rather than code
    • Roots in my 2021 deck, The Future of UX and AI-driven Opportunities

    11named measures, from intent defect rate to energy per outcome

  • 2025 – 2026

    Comprehension, the missing layer of DPI

    The grand picture

    Comprehension, not access and not enrolment, is the true unit of digital inclusion. India built identity and payment rails. Next it must build the comprehension layer.

    The legs
    • Kya Likha Hai (KLH): a protocol that explains the consequence of a high-stakes act by voice, in the person’s language, before the act
    • The comprehension artefact: a record of understanding beside the record of consent, with a teach-back check
    • Five working instruments in the browser: Claude does the reasoning, Sarvam does the Indian-language voice
    • Speaking Digital, a book manuscript of 92,000 words

All work

Twenty minutes is enough to tell whether I can help.

Tell me the decision in front of you. If it is mine to help with, we scope a fixed-fee engagement. If it is not, I will say so and suggest who can.