The profile
Fifteen years building and leading innovation and digital transformation inside government, paired with an active AI practice building agentic systems for senior public-sector audiences.
Fredericton, Canada · Atlantic Time · Long-running work with teams in Ottawa, Abu Dhabi, and Dubai
Exhibit A · The operating environment
Forty-four role-scoped agents, organised the way a firm is: one router at the hub and forty-three specialists across eight clusters. Explicit handoffs, quality gates, and an advisor council whose standing job is to disagree. Hover or tap any role to read its remit.
This page was produced inside it.
Exhibit B · The governance stance
One of your questions is how a country keeps public-sector AI accountable and sovereign. Here is the rule I build to. The AI reads the evidence and recommends. It never makes the call. A named person makes each decision, and that person is the one who is accountable for it.
The system is built so the AI cannot write a final decision, only its advice. Every time it runs, it records what it recommended, on what evidence, and with which model. So later, a manager or an auditor can confirm that a person decided, and see exactly what the AI put in front of them. The accountability still rests with the person. What the data model adds is proof: it makes that accountability easy to check.
This rule comes from practice. I have spent nine years assessing Dubai's government against its own standards, and I build agentic systems for that world: an assessment platform, an AI jury of six declared perspectives that deliberates in front of senior officials, and a policy simulator. I published the full framework, including where it fails, at agentic.shiftflow.ca.
Exhibit C · Shipped work
Built tools and platforms from real public-sector practice. Each card names the CDPI question it speaks to.
AI deliberation instrument
A facilitated sensemaking instrument for public-sector transformation. Frame a real organizational challenge, draw a spread from five suits (barriers, enablers, theories, tools, provocations), and a panel of six AI personas deliberates it live, streaming to a synthesis of insights, actions, and risks. The web app fronts a 110-card physical deck used in real rooms.
live · no sign-in
Multiplayer workshop game
A multiplayer game that names organizational dysfunction so a room can finally talk about it. A facilitator opens a room with a code and a QR, up to twelve players join on their own phones, and the post-game pattern analysis is the actual workshop artifact. Seven decks carry 172 prompts and 413 responses.
live · public
AI advisory simulator
An AI board of advisors. Set your context, seat up to three advisors from themed boards, pose a strategic question, and watch a streamed session render advisor rounds, a synthesis, and a tension map of where the board disagrees. Variety and tension are where innovation lies.
live · public
Futures imagination · AI imagery
Stuart Candy and Jeff Watson's futures game, digitized with attribution (CC BY-NC-SA). Draw Arc, Terrain, Object, and Mood cards, write the artifact that would exist in that future, and render it with AI imagery. Built for public-sector foresight sessions where imagination needs structure.
live · no sign-in
Strategic futures · AI-assisted
Bill Sharpe's Three Horizons framework as a guided canvas. Work through the present system, the future vision, and the bridge between them; Claude reads the whole canvas and returns a what / so-what / now-what analysis with two deliberately uncomfortable 90-day commitments, rendered as a printable horizons map.
live · no sign-in
Systems model · AI-assisted
The Berkana Institute's Two Loops model of system change as a reflective instrument. Eleven prompts across the dominant system, the transition, and what is emerging; Claude headlines each answer and annotates a living map of the user's own system in transition.
live · no sign-in
Podcast & wisdom library
Conversations about the how of public innovation, co-hosted with Jason Pearman. Twenty-one episodes with edited transcripts, auto-derived tables of contents, and host marginalia. The editorial pipeline behind every reading room is an AI-assisted practice with a human verify gate on every ship.
live · public
Three positions
How do we drive growth of all the layers of the AI stack industry in a country?
Growing a country's AI industry starts with people and institutions. The models get better and cheaper every year. The scarce thing is a public sector that knows how to run AI, check its work, and change course when it fails. Build that skill and the country captures value at every layer of the stack. Skip it and the value leaks out, no matter how many models get bought.
How does a country balance sovereignty in public procurement of AI with access to frontier models?
Sovereignty and frontier access look like a trade-off. They stop being one when the country controls the layer where decisions happen. Keep three things in public hands: the gateway every AI request passes through, the record of what the AI did and why, and the rule that only a named person can make a final decision. Behind that boundary, a government can use a frontier model today and a home-built model next year, and swap between them without giving up control. As my published framework puts it, control gets enforced by how and where the system runs, not by policy alone.
Where will most value in the AI value chain be derived from, and how should this be captured?
India already showed where the value goes: it grows on top of open rails. Identity and payments became public infrastructure, and a whole application economy grew above them. The same move is open for AI. Make the base layers shared and public, things like registries, open evaluation, and public compute, and let builders compete above them. The countries that lay those rails first will host the value that grows on top. The Global South can move first, on purpose.
The contribution
I read the DPI Wiki and CDPI's program pages before writing this. Four contributions, each mapped to work CDPI has already named.
DaaS bootcamps & co-creation
CDPI moves countries from planning to implementation with bootcamps and co-creation workshops. That is capability building, and it is the core of my record: I directed a 30-person team delivering digital capability programs to thousands of public servants at Canada's Digital Academy, built New Brunswick's public innovation function from the ground up, and designed the lab methodology and training program inside Abu Dhabi's customer-experience transformation. I would design and run the institutional side of DaaS: the sessions, the curriculum, and the follow-through that turn a deployment into a working habit.
The AI Assistant offering
CDPI's upcoming DaaS cohort includes an AI Assistant for citizen-facing information. I build agentic systems for senior public-sector audiences, and I published the framework I build to, including where it fails. The rule in the governance section above travels well: the AI recommends, a named person decides, and the record proves it. I would help shape the reference architectures, procurement language, and evaluation practice so countries adopt AI assistants they can audit and steer.
Decision frameworks & assessment
The wiki already helps a country answer "Is my system a DPI?". Countries now need the same instrument for AI: is the system accountable, is the capability real, is the institution ready to run it. I have spent nine years assessing Dubai's government against its own standards and delivered a CX maturity assessment alongside international assessors. I would draft the AI-readiness assessment in the same house style: plain questions, defensible scoring, and an honest path for countries that should wait.
Country engagements
CDPI lists eight country engagements today, from Brazil and Perú to Cambodia, and each one runs through the same hard step: getting a minister's office, an engineering team, and local industry to a shared decision. Colleagues describe that bridge work as my defining strength, and I have turned partnerships across three orders of government, employers, and community organizations into results. I also sit on the board of Code for Canada, watching a G7 country pay the price of fragmented digital infrastructure. I would carry stakeholder alignment and first-use-case work inside engagements.
Program names and engagement counts above come from cdpi.dev and the DPI Wiki, read in August 2026.
The letter
August 2026
Hiring Team
Centre for Digital Public Infrastructure
Re: AI Associate
Dear Colleagues,
The Centre for Digital Public Infrastructure occupies a particular place in a moment that will not stay open long. Countries are deciding now how artificial intelligence enters public service, and most of those decisions will harden into architecture within a few years. What follows from them is consequential: whether a government can change its mind later, whether citizens can see how a decision about them was made, and whether the capability a country builds belongs to it. That work takes deliberate design. It happens because someone specifies the interface, holds the standard, and builds the capability to use it. I am writing to express my interest in the AI Associate role because I believe in that work and have spent my career in service of it.
I understand the task ahead. CDPI supports the India AI Mission and works with governments building AI capacity, which means translating principle into procurement language, reference implementations, and guidance that officers will implement. I have led that kind of conversion before. As Executive Director of Open Government and Innovation in New Brunswick's Executive Council Office, I managed a $2.2 million budget and turned a government priority into a standing portfolio, including a public innovation framework, the Deputy Minister Innovation Council, the Policy Innovation Challenge, and the platform that carried knowledge management and impact measurement. As Director and Executive Lead at the Government of Canada's Digital Academy, I directed 30 staff across five teams and delivered capability programs to thousands of public servants.
This role calls for someone who can serve as a credible bridge across government, industry, and the technical community. Colleagues describe that as my defining strength: the ability to bring a deputy minister, an engineer, and an industry executive to the same table and move them toward a shared decision. I founded NouLAB, Atlantic Canada's public and social innovation lab, to do exactly that, and I have turned partnerships across three orders of government, employers, and community organizations into results.
My grounding in digital public infrastructure runs through its absence. I serve on the board of Code for Canada, the national organization now building Canada's DPI and confronting the incompatible provincial digital identity systems fragmenting the country. Canada introduced digital identity in 2012 and now trails 37 countries. It trails more than 70 on open banking and real-time payments. I spent a decade inside the institutions responsible, and money never was the constraint. Canada failed because no institution held mandated authority over the interface, and jurisdictions stayed free to fragment.
I am further into a career than this posting describes, and I have weighed that carefully. I am applying because the problem interests me more than the title does. I also have no India experience, and reading about the India Stack teaches nobody how institutions actually move.
The decisions countries make now about AI in public service will hold for a generation. I would be honoured to bring my experience in public-sector transformation, procurement, and capability building to that work, and I welcome the opportunity to discuss it. Thank you for your consideration.
Sincerely, Nicholas Scott
Curriculum vitae
Public-sector transformation leader and professional facilitator with fifteen years building and leading innovation and digital transformation inside government, now paired with an active AI practice building agentic systems for senior public-sector audiences.
Boutique innovation-and-transformation practice, now paired with an active AI practice: a published agentic-government framework, a suite of agentic prototypes for senior officials, and two products.
Abu Dhabi's Effortless Customer Experience transformation: a Service Innovation Lab methodology, a multi-level CX program, and a CX maturity assessment with international assessors.
Assess entries against the smart-government standard with site visits, interviews, and written scoring; three years inside the 360 cycles alongside senior Dubai officials.
Helping organizations adapt amid digital disruption with agile, user-centred approaches in service of the public good.
Executive lead at the Digital Academy, leading a 30-person team across the Digital Accelerator Programme, UX research, digital experimentation, and the vExpo platform.
Built a public-sector innovation function from the ground up; established the Public Innovation Challenge, the DM Innovation Council, an internship program, and the GNB Innovation Lab; reached 500+ staff.
Co-founded New Brunswick's first public and social innovation lab; built the NouLAB Academy with MaRS and GovLab NY; ran five annual lab programs and the Economic Immigration Lab.
Networked-governance approach to evidence-based policy across 1,000+ members; established the GovMaker Conference and hosted the Canadian Open Data Summit.
Courses on leadership for social innovation, technology and social change, and the social organization of the internet.
Education
Training
Invited speaking
Invited to speak on innovation, digital transformation, and public-sector reform across four continents.
Contact
Everything above is live if you want a closer look.
Designed and built in one working day, inside the operating environment it describes.
No tracking on this page. Not indexed.