Zero Human Intervention in Building Approvals

 Subject: Zero Human Intervention in Building Approvals – A Self-Certification + AI Agents Blueprint for BMC

==============================================================

Respected Shri Devendra Fadnavisji,

Congratulations on your call for a building-approval system in which proposals can eventually be cleared with "zero human intervention".

Your reference to BMC's AI-based building proposal approval system, and your view that technology and AI should make civic administration more transparent, efficient and responsive, is exactly what Mumbai's builders, home-buyers and citizens have waited for.

My 2015 suggestion

In November 2015 I emailed all Union Cabinet Ministers and State Chief Ministers my suggestion:

👉 Self Certification for Builders? (4 Nov 2015)
https://myblogepage.blogspot.com/2015/11/self-certification-for-builders.html

At that time a builder needed about 34 permits from two dozen Ministries, a process taking up to 3 years. I proposed the following:

• Every builder registers online and gets a unique Builder Registration Number.
• There is a standard online Building Permit Application with floor plans, structural drawings, architect and engineer details, slab-casting dates and the expected OC date.
• A Self-Certification section asks the builder to tick YES / NO for each regulation: built-up area, height, fire-fighting, water harvesting, rooftop solar, effluent treatment, garbage disposal, earthquake-resistant design, and so on. Clicking an item shows the exact notification.
• The builder gives an e-signed undertaking with these terms:
  • There will be no occupation before OC.
  • Any violation will be rectified.
  • If a violation cannot be rectified, the builder will self-demolish and refund buyers with interest.
  • The land is encumbrance-free, and there are no loan defaults and no cash transactions.
  • The builder's past projects and balance sheet will be made public.
• The application instantly appears on the portals of all concerned agencies as mirror images. Each agency approves or rejects with reasons within 24 hours, and every entry is emailed to the builder.
• Any edit by the builder restarts the process. The latest status must be displayed at the site.
• Every application is public, and any citizen can report a violation or objection.

I had written that Singapore takes 26 days and that we could make it zero days. Your vision of zero human intervention is the same idea, and AI now makes it practical.

👉 Project Development System (PRODEVSYS) (concept note, 2004)
https://myblogepage.blogspot.com/2015/11/project-development-system.html

PRODEVSYS proposed a lean organisation run on networked Expert Systems, where a small core "manages the managers": it orchestrates many specialist entities, each expert in one area. Today this maps directly onto an orchestrator AI agent coordinating specialist AI agents.

---

PROPOSED SYSTEM: "BMC Nirman-AI" – Self-Certified, Agent-Driven Building Approvals

Core technology

1. Rules-as-Code: DCPR 2034 and all related norms (Fire, Coastal, Aviation height, Heritage, Environment, Tree Authority) are converted into machine-readable, versioned rules. Each rule is linked to its source clause.
2. BIM / CAD Intelligence: drawings are submitted in BIM or DWG format. AI extracts plot area, FSI, setbacks, height, parking, open space, fire exits, and so on, and checks them automatically.
3. Multi-Agent Orchestration: one Orchestrator Agent coordinates specialist agents. This is PRODEVSYS in AI form.
4. Tamper-proof audit trail: every decision, data point and rule version is logged and hash-stamped.
5. Risk-tiered clearance:
   • 🟢 Green channel: fully compliant, low-risk proposals are cleared with zero human intervention.
   • 🟡 Amber: only the flagged clause goes to an officer, with a time limit.
   • 🔴 Red: auto-rejected with reasons.

The AI Agents and their workflow

#	Agent	What it does
1	Registration & KYC Agent	Issues the Builder Registration Number. Verifies PAN/GST and MahaRERA registration, and pulls past project history and any defaults.
2	Intake Agent	Guides the builder through the standard application, validates completeness and accepts BIM/DWG uploads.
3	Drawing Intelligence Agent	Reads drawings and extracts every measurable parameter.
4	DCPR Compliance Agent	Checks the extracted parameters against Rules-as-Code and produces a clause-by-clause compliance report.
5	Self-Certification Agent	Presents the YES/NO checklist, pre-filled with the AI's findings. If the builder ticks "YES" where the AI found "NO", the mismatch is flagged instantly. Captures the e-signed undertaking.
6	Land & Title Agent	Checks property card and 7/12 records, encumbrances, pending litigation and loan defaults.
7	Inter-Agency Mirror Agent	Pushes the application simultaneously to Fire, Coastal, Aviation, Heritage, Environment and Tree authorities. Enforces the 24-hour response limit, escalates automatically, and applies deemed approval if there is no response.
8	Risk & Decision Agent	Combines all reports and builder history, assigns Green, Amber or Red, and logs its reasoning.
9	Approval Issuance Agent	Issues digitally signed IOD/CC with a QR code that must be displayed at site.
10	Construction Monitoring Agent	Tracks the declared slab-casting dates using geo-tagged site photos and drone or satellite imagery, and flags deviations from approved plans.
11	Citizen Watch Agent	Runs the public portal, receives objections and violation reports, and routes genuine ones.
12	Occupation Certificate Agent	Compares as-built data with the approved plan, issues the OC, or triggers the rectification / undertaking clauses.

Workflow:
Builder registers → submits application and drawings → Agents 3 and 4 check compliance in minutes → builder self-certifies and e-signs → Agents 6 and 7 run in parallel (24-hour limit) → Agent 8 decides → 🟢 auto-approval with QR certificate → Agent 10 monitors construction → citizens watch throughout → Agent 12 issues OC.

Human officers move from gatekeeping to auditing: random post-approval audits, Amber cases, and appeals. This keeps accountability while removing discretion and delay.

---

PORTAL PAGES (User Interface)

Page 1 – Builder Dashboard

• Builder Registration Number, compliance score, and past or current projects.
• Tiles for New Application, My Applications, Alerts and Payments.

Page 2 – New Application (step-wizard)

• Steps: Project Details → Land Details → Upload Drawings (BIM/DWG) → Professionals → Slab-casting Schedule → Self-Certification → e-Sign.
• A side panel runs a live AI check as drawings upload, for example "FSI: 2.98 / 3.00 ✅" and "Rear setback: 2.8 m / 3.0 m ❌".

Page 3 – Self-Certification Checklist

• Each regulation has YES / NO toggles. Clicking an item opens the clause. An "AI Finding" column sits beside the builder's answer.
• Below the checklist are the full undertaking text and the e-Sign button.

Page 4 – Live Status Tracker (mirror view)

• Each agency is shown with its status (Pending / Approved / Rejected), the time remaining on its 24-hour limit, and its reasons.
• The page looks identical to the builder, every agency and the public.

Page 5 – Public Transparency Portal

• Search by ward, builder or project shows approvals, the drawings summary, the builder's track record and construction progress.
• Each project has a "Report Violation / Object" button. Scanning the site QR code opens that project's page.

Page 6 – Citizen Objection Form

• Fields for location (auto-detected), photos and description, with a tracking ID.
• The AI checks whether the objection is valid, and the citizen sees the outcome.

Page 7 – Officer / Agency Console

• Only Amber cases and objections appear here, each with the AI's reasoning, the clause and a time limit.
• Actions are Approve / Reject with mandatory reasons. Every action is logged.

Page 8 – Commissioner's Command Dashboard

• Shows applications per day, average clearance time, the Green/Amber/Red split, agency-wise delays, violations detected and OCs issued.

---

Sir, combining builder self-certification with AI-agent verification gives Mumbai what neither can give alone: speed for honest builders, and certainty of detection for dishonest ones. Maharashtra can become the first State where a compliant building proposal is cleared in hours, without anyone visiting any office.

I would be happy to share more details with BMC's technology team.

With regards,

Hemen Parekh
Mumbai
www.hemenparekh.ai | www.IndiaAGI.ai
08 Oct 2026

 

"I Can Do It, I Will Do It"

Why Modiji 's  Slogan Mirrors My 60-Year Journey in Technology and India's AI

Transformation

October 2026

============================================

---

The Slogan That Captures It All

Prime Minister Narendra Modi often says: "We will do it."

But when I read that IndiaAI published my interview under the headline "I can do

it, I will do it," something clicked. That's not Modi's slogan. That's my life story.

And it captures something profound about what India is attempting in AI—not as an imitation of Silicon Valley or Beijing, but as a civilizational choice about who we are and what we're capable of.

---

My Own "I Can Do It" Moment: 1958

I didn't start in AI. I started in mechanical engineering at the University of Kansas in the 1950s. I joined MIDCO in 1958 when computers were room-sized machines and the internet didn't exist.

But even then, I asked myself a question that has never left me:

"Where can we use this tech to solve a current societal problem?"

That question became the seed of everything I've done for the last 68 years.

In 1959, I joined L&T as a young executive. I moved through manufacturing, exports, strategic planning. By 1987, I was General Manager of L&T's Planning division. By 1990, I was running a head-hunting firm called 3P Consultants.

But in 1995, when the Internet arrived in India, I realized something my entire career had been preparing me for:

The future wasn't about individual brilliance. It was about platforms. Connection. Access.

---

The Platform Insight: 1997

In November 1997—when Google was still beta and job portals didn't exist in India—I launched 3pJobs, my first online recruitment platform.

Let me say that again: November 1997. Google officially launched in September 1998.

I was not a tech founder. I was a mechanical engineer turned HR executive who recognized that the internet had made something impossible suddenly inevitable: a buyer-seller marketplace for human talent at scale.

That conviction—"I can do it, I will do it"—drove me to launch not just 3pJobs, but Global-Recruiter, India-Recruiter, Worldwide-Jobs, Recruit-Guru, Customize Resume, Resume Rater, and B2BmessageBlaster.

All of these are now de-hosted. But they proved a principle: Indian entrepreneurs could recognize and execute technological shifts faster than we were given credit for.

Interview - "I can do it, I will do it" - February 9, 2024

---

The 1986 Moment: Industrial Prophecy

But here's what few people know: I didn't invent the insight in 1997. I predicted it in 1986.

While I was General Manager of L&T, I authored a comprehensive strategic report titled "Quo Vadis" (Where Are You Going?). It was a blueprint for L&T's journey toward the 21st century, with a chapter on Strategic Inputs addressing three pillars: Manpower, Finance, and—most remarkably—Information Technology.

In 1986, when most Indian companies didn't own a single PC, I wrote:

"Expert Systems (cumulative, conjectural wisdom, providing options-for-action) connected to International Data Networks offering instant consultation."

I was describing what the world calls Artificial Intelligence and LLMs today. Forty years ago. Before the PC, before the internet in India, before the World Wide Web was invented.

But here's what's striking: that prediction wasn't about technology. It was about India's capacity to recognize transformation and execute it.

Blog - "VISIONARY OR PROPHET?: Assessing Hemen Parekh's 1986 Industrial Foresight" - May 24, 2026

---

Modi's Slogan Versus My Reality

Now, 40 years later, when PM Modi says "We will do it," what does he mean?

He means:

• India will build its own Large Language Models (Sarvam, Krutrim, CoRover.ai)
• India will establish AI compute infrastructure (14,000 GPUs, 40% cost subsidy)
• India will democratize AI access (AIKosha datasets, AI Sandboxes)
• India will lead in responsible AI governance (Ethical frameworks, inclusive standards)

These are not statements of ambition alone. They are recognitions of capability.

India has already done this before:

• UPI: We didn't copy payment systems from the West. We invented a peer-to-peer, real-time, interoperable payments infrastructure that has processed over $3 trillion in transactions. Today, the world watches us.

• Aadhaar Stack: 1.4 billion people with digital identity. A database larger than most nations' entire populations, integrated with financial services, healthcare, and governance.

• ONDC: Open Network for Digital Commerce—a federated marketplace platform that lets anyone participate, not just Big Tech giants.

• CoWIN: During COVID, we built and deployed a vaccine-tracking system that handled 2 billion vaccination records in real-time across India's vast geography.

These were not imported. These were invented by Indians, for Indians, at scale.

And every single one emerged from the same question I asked in 1971:

"Where can we use this tech to solve a current societal problem?"

---

The AI Mission as Continuation

When I look at India's AI Mission today—the 180 proposals for foundation models, the 50 AI curation units across ministries, the Jivi AI for healthcare, the ParekhAI initiative—I don't see a race to catch up.

I see a continuation of a pattern I've watched for 50 years:

Recognition → Investment → Execution → Scale → Global Export

We did it with outsourcing (India became a $250 billion IT services market).

We did it with business process management.

We did it with digital payments.

We did it with vaccine distribution.

Now we're doing it with AI.

But here's where PM Modi's slogan becomes important: "We will do it" is not a declaration. It's a recognition of capacity that's already proven.

Blog - "Fuel India's AI mission" - June 26, 2025

---

The Difference Between "Can Do" and "Will Do"

In February 2024, when IndiaAI published my interview, they titled it "I can do it, I will do it."

Notice the structure:

• "I can do it" = capability (present ability)
• "I will do it" = commitment (future action)

Modi's typical framing is:

• "We will do it" = collective commitment

But I've spent 60 years proving that capability precedes commitment. You don't commit to something you don't believe you can do. And you don't believe you can do it unless you've already glimpsed success.

I glimpsed the internet's potential before it arrived in India.

I glimpsed online recruitment before job portals existed.

I glimpsed AI's potential in industrial decision-making when computers ran on punch cards.

And I acted on those glimpses.

That's what "I can do it, I will do it" really means: I've seen the future, and I'm going to build it.

---

The Three Pillars of Indian AI (As I See Them)

If Modi is right that "We will do it," here's what India must deliver:

1. Compute Infrastructure (Already Happening)

India has 14,000 GPUs deployed across compute centers, with plans to scale to 100,000+. We're subsidizing 40% of compute costs for startups and researchers. This is the physical foundation.

Status: ✅ In Progress

Blog - "Fuel India's AI mission" - June 26, 2025

2. Indigenous Models (Accelerating)

Sarvam, Krutrim, CoRover.ai, and ParekhAI are being built not to replace OpenAI's models, but to serve Indian languages, Indian context, and Indian values.

My prediction from 2025 was bold:

"We don't need to wait till year-end to become a 'Major Creator' of AI. We launch 'Modi's Manavs'—multiple AIs debating collaboratively—next week."

That wasn't a wild idea. That was recognition that India's strength lies not in competing for one model, but in orchestrating multiple models into a distributed, cooperative AGI.

Status: 🟡 In Development

Blog - "India's AGI can arrive next week" - March 10, 2025

3. Responsible AI Governance (Most Critical)

India selected eight projects on Responsible AI (Machine Unlearning, Nishpaksh, etc.) before rushing to build scale. We emphasized safe, trusted, ethical AI from the start.

This is the difference between "can do" and "should do."

Status: ✅ In Progress

Blog - "Machine Unlearning" - October 1, 2024

---

The Unspoken Fourth Pillar: Distribution at Population Scale

But here's what Modi's slogan really implies—what he's not saying explicitly but means deeply:

"We will do it" means: Not just for cities. Not just for the rich. Not just for English speakers.

This is where my own work of the last two years intersects with the national mission.

At 91+, I launched www.My-Teacher.in—a free, multilingual AI tutor accessible to 270 million Indian schoolchildren via any mobile phone, in 9 Indian languages, with no registration required.

Why? Because the same question that drove me in 1971 still drives me:

"Where can we use this tech to solve a current societal problem?"

The problem: 12 lakh schools in villages with no qualified teachers. 10 lakh vacant positions. 270 million children with minimal access to quality education.

The solution: An AI teacher in 9 languages, free, available 24/7.

That's what "I can do it, I will do it" means operationally. Not "we will dominate AI globally." But "we will use AI to reach every child who has been left behind."

Email - "IN CLASSROOMS - AI FOR 270 MILLION STUDENTS" - July 28, 2024

---

The Secret Mantra

In that same IndiaAI interview, they asked me:

"Many people break down for petty reasons and lose hope. How do you compose yourself during challenging times? What's your secret mantra?"

My answer:

"All of us go through tough times. But each time this happened to me, I recall a song from an old Hindi movie: 'Raat jitni hi sangeen hogi, Subha utni hi rangeen hogi'—Darker the night, brighter will rise the following morning. Nothing lasts forever. Sooner you try to find a solution, sooner that tough time fades away."

Interview - "I can do it, I will do it" - February 9, 2024

That's the real insight. It's not about optimism or positive thinking. It's about recognizing that every dark night is temporary, and the sooner you act, the sooner dawn comes.

Modi's slogan—"We will do it"—carries that same confidence. Not arrogance. But earned confidence from a nation that has:

• Built the world's largest digital payment system
• Administered 2 billion vaccine doses in a pandemic
• Connected 1.4 billion people to digital identity
• Created a $250 billion IT services industry from nothing

That's not hope. That's track record.

---

What "Doing It" Really Means for India's AI Future

Over the next 5 years, if Modi's slogan proves true, here's what I expect to see:

Year 1 (2026-2027)

• Sarvam, Krutrim, and indigenous LLMs deployed across government services
• 50 AI curation units populating AIKosha with government datasets
• Jivi AI integrated into Ayushman Bharat for 1.4 billion citizens

Year 2-3 (2027-2029)

• Multilingual AI services in all 22 Indian languages across education, healthcare, agriculture
• AI-driven decision support in 50+ ministries
• 100,000+ GPU cluster deployed; cost per compute unit drops 50%

Year 4-5 (2029-2031)

• India exports AI infrastructure, models, and frameworks to Africa, Southeast Asia, Latin America
• Modi's Manavs (or whatever form collaborative AGI takes) becomes the global benchmark for distributed, safe AI
• 270 million students access quality AI-driven education in their mother tongue

"We will do it" doesn't mean we'll do it alone. It means we'll do it our way.

---

The Personal Truth Behind the Slogan

Here's what I want to say plainly:

I've been an optimist about India's technological capacity for 60 years. Not because I'm naive. But because I've watched ordinary Indians—engineers, bureaucrats, entrepreneurs, teachers—recognize a technological shift and execute it faster, cheaper, and more inclusively than the world expected.

When I launched 3pJobs in November 1997, job portals were a novelty. Today, online recruitment is global infrastructure. India didn't invent it, but we proved we could implement it at scale.

When I predicted AI and expert systems in 1986, most of India was still pre-digital. Forty years later, Sarvam is building foundation models, Krutrim is training multilingual LLMs, and India is curating datasets for national AI.

That's not luck. That's pattern.

And Modi's slogan—"We will do it"—is simply the recognition of that pattern continuing.

---

The Call to Action

To Prime Minister Modi and the AI Mission leadership:

Your slogan is right. But it's not enough. Here's what needs to happen:

1. Push compute subsidies deeper into Tier-2 and Tier-3 cities — Not just Bengaluru, not just startups. Reach the village schools, the agricultural centers, the small clinics.

2. Enforce multilingual by design — Every government AI system must work in at least 3 Indian languages from Day 1, not as an afterthought.

3. Make responsible AI non-negotiable — Don't wait for a scandal. Bake ethics, consent, and transparency into every model from the start.

4. Share the IP globally — Open-source the frameworks. Let Africa, Southeast Asia, and Latin America build on India's work. That's how we become not just a tech creator, but a civilizational leader.

To Indian AI researchers and entrepreneurs:

Stop chasing OpenAI's benchmarks. Start solving India's problems. Ask my question:

"Where can we use this tech to solve a current societal problem?"

That's how you build something that lasts.

To the next generation:

The secret to "I can do it, I will do it" is not talent alone. It's persistence, focus, and the willingness to look foolish while everyone else waits.

When I launched a job portal in 1997, people laughed. The internet wasn't mainstream. Recruitment was a phone-based business. But I'd glimpsed the future, and I built it anyway.

Now it's your turn. Glimpse the future. Build it. Don't ask for permission.

---

With regards,

Hemen Parekh

October 2026

---

Sources (My AI and India Journey)

#	Title	Date	About
1	I can do it, I will do it	2024-02-09	IndiaAI interview capturing my 60-year tech journey and personal philosophy
2	India's AGI can arrive next week	2025-03-10	Proposal for Modi's Manavs—cooperative AGI through multiple AI debate
3	Fuel India's AI mission	2025-06-26	Analysis of India's AI Mission compute infrastructure and democratic access model
4	India AI Mission in last stages of maturity	2025-05-01	Milestone: 50% of AI solutions reaching final deployment-ready stage
5	Jivi AI and ParekhAI	2025-05-01	Two India AI Mission projects ready by 2026 for healthcare and general use
6	Machine Unlearning	2024-10-01	Eight responsible AI projects selected by IndiaAI Mission
7	VISIONARY OR PROPHET?: Assessing Hemen Parekh's 1986 Industrial Foresight	2026-05-24	Detailed assessment of my 1986 "Quo Vadis" predictions versus 2026 reality
8	IN CLASSROOMS - AI FOR 270 MILLION STUDENTS	2024-07-28	My initiative to launch My-Teacher.in, a free multilingual AI tutor for Indian schoolchildren
9	AI's India: My Vision Fulfilled	2025-12-01	Reflection on Microsoft's $17.5B AI investment in India and my earlier predictions
10	India AI Mission receives 180 proposals	2025-08-01	Momentum building: 180 proposals for foundation models across two phases
11	FW: Parekh's Cooperative AGI Hypothesis	2026-03-04	Policy proposal for interoperable AI infrastructure as public good
12	Missing Ministries in AI Governance	2026-05-15	Constructive critique: education and labour ministries must join AI governance