About
Built in the open, roundtable by roundtable.
IIDA is an initiative of ABAIA, the All Bharat Artificial Intelligence Association — a non-profit, non-political body building a support system for builders, innovators and industry leaders. The question that started it was blunt: AI is enormous, so why is our city behind?
The record
Roundtables
First roundtable
Strategic framework for deeptech AI innovation in India. The founding argument gets its first hearing.
Policy session
AI industry diffusion compared across two states, and what an alliance can do that a department cannot.
Second roundtable
Streams and leads formalised. The work stops being a discussion and starts having owners.
Third roundtable
Vision and mission ratified. FRO framing settled around SMEs and export.
Sessions run 90 to 120 minutes with 25 to 30 participants, closed room. Not a conference, not a panel, not a networking event — a working group that produces a plan someone owns by the end of it.
What could go wrong
Risk register
Eight ways this fails.
Written before starting, published while it is still awkward to publish.
Credibility cold start
No track record yet. Funders and researchers both wait for first movers.
Move
Anchor advisor first, entity second. Ship one visible artefact within 90 days.
Scientific leadership
Finding a Director with research credibility, operating ability and a volunteer ethos.
Move
Recruit the Scientific Director in the open. The role is posted, not filled.
Volunteer reliability
Sustained quality from unpaid contributors requires real structure, not goodwill.
Move
Tiered onboarding — start with 20 chosen contributors, not 200. Two-strike norm on missed deliverables.
Quality control
Distributed research drifts to mediocrity without rigorous review.
Move
Low-originality work exits to a blog or whitepaper rather than being carried as research.
Funding before proof
Indian funders want output first. Output needs funding. Chicken and egg.
Move
Compute-first fundraising — easier to win, and it demonstrates resourcefulness.
Compute access
Frontier work needs GPUs that volunteers cannot personally provide.
Move
Compute credits as a patron contribution class, equal in standing to capital.
Focus discipline
Four exciting tracks plus eager volunteers makes scope creep the default outcome.
Move
Kill criteria written before work starts. One focused mission committed by Q4 of Year 1.
Retention against industry
Top contributors get poached. We cannot win on salary alone.
Move
Authorship by CRediT taxonomy, public contribution tracker, and a real path from volunteer to paid core.
Where things stand
Status
Honest status, since you asked.
- The legal entity is being established. We are not yet a registered Section 8 company and do not claim 12A, 80G or FCRA status.
- No Scientific Director has been appointed. The role is open and being recruited in public.
- No paper has been published. The first artefact is the near-term priority, ahead of fundraising.
- All four research tracks are seeking leads.