DPI Deep Dive — Saturday | September 19, 2026

Layer 6 — Governance & Grievance — is where India’s DPI stack talks back: the layer that registers the citizen’s complaint, routes it, clocks it and closes it. This week it made news four times over, and the four windows fit together. On September 16, the government switched on the dedicated web portal for Special Campaign 6.0, the annual, government-wide pendency purge. ^{1 2} On the numbers side, DARPG released the 51st CPGRAMS monthly report showing 2,16,827 grievances redressed by Central ministries in July, ³ while the daily bulletins recorded 13,557 disposals on September 17 alone. ⁴ On September 18, DARPG ran a capacity-building workshop on special e-governance projects for the North-Eastern states and Ladakh. ⁵ And on September 17, the Ministry of Minority Affairs convened a workshop on AI-enabled governance built on Sarvam AI’s Indus, the same week the grievance layer’s own AI assistant — Samadhan Didi — continues to scale on Bhashini. ^{6 7} The through-line: the layer has finished solving access (anyone can file, in any scheduled language) and is now industrialising disposal — with AI moving into the triage seat. The binding constraint is shifting from “can you file?” to “does closure mean relief?”

1. Special Campaign 6.0: The Grievance System’s Annual Janitor, Now on a Dashboard

Union Minister Dr. Jitendra Singh launched the dedicated web portal for Special Campaign 6.0 at the National Media Centre on September 16, before nodal officers drawn from over 80 ministries and departments — the campaign’s sixth edition since DARPG became its nodal coordinator in 2021. ^{1 2} The architecture is familiar: a preparatory phase from September 16 to 30, in which each ministry identifies its pendency targets and cleanliness sites, followed by an implementation phase from October 2 to 31 in which the targets must be met and reported weekly on the portal. ² Four pillars have defined the campaign since inception — swachhata (cleanliness), physical record weeding, space reclamation through scrap disposal, and reduction of pendency in public grievances, appeals and parliamentary references. The week’s prep-phase releases show the machinery engaging across government: the Ministry of Mines is highlighting scrap and e-waste disposal, ⁸ the Department of Expenditure and the Health Ministry have issued their campaign plans, and the Panchayati Raj Ministry — whose secretariat already runs fully on eOffice — has confirmed participation. ⁹

The scoreboard from Special Campaign 5.0 explains why ministries take this seriously: as on November 11, 2025, the campaign had covered 11.6 lakh office spaces, disposed 7.36 lakh public grievances, earned ₹816.75 crore from scrap disposal, and freed 233.7 lakh square feet of office space. ¹⁰ The consumer read is double-edged. Campaign mode genuinely works as a scheduling device — October is reliably when file backlogs and grievance pendency actually clear, because a dashboard, a deadline and a nodal-officer meeting create the one thing ordinary administration cannot: urgency. But campaign mode also has a known failure mode: disposals optimised for the count, not the citizen. The 21-day norm under the August 2024 comprehensive guidelines is the baseline the campaign must not distort — a grievance closed hastily on October 30 to make the dashboard is worse than one closed properly on November 15. The concrete ask: the new portal should publish closure-quality — complainant feedback ratings and appeal rates — per ministry alongside disposal counts. A dashboard that shows only disposals measures the janitor, not the relief.

The cross-layer angle: Special Campaign is Layer 6 turning inward on its own plumbing — eOffice fileless operations and record management (the layer’s internal rails) are simultaneously its tool and its object. Panchayati Raj’s “fully eOffice” secretariat is where every ministry is headed, and the campaign is the forcing function.

2. CPGRAMS by the Numbers: A Disposal Machine, and Its Five-Lakh Caveat

The platform’s steady-state output is now enormous and worth stating plainly. The 51st CPGRAMS monthly report records 2,16,827 grievances redressed by Central ministries and departments in July 2026 — extending a streak of more than a lakh Central disposals every month that is now in its fourth year. ³ DARPG also released the 48th monthly report for States/UTs for July, the companion volume that tracks the state tier where the heaviest backlogs sit. ⁴ The daily drumbeat continues underneath: 13,557 grievances disposed across the Centre and states on September 17, ⁴ 17,916 on September 10, ¹¹ and 16,630 on September 8. ¹² For 2026 as a whole through July 15, over 15.2 lakh grievances were received, with the average Central disposal time at 13 days — against the 21-day norm, and against 157 days in 2014. ^{10 13}

But the most important number in the layer this year is one DARPG itself surfaced: an analysis of more than 5 lakh recurring grievances across four years showing that fast closure does not guarantee relief — with PM-KISAN payment failures, income-tax refunds and banking-fraud resolution the persistent high-volume categories. ⁸ The systemic read: CPGRAMS has two accountability regimes, and only one of them is measured. Speed is measured exquisitely — 13-day averages, 46-month disposal streaks, daily bulletins. Outcomes are measured thinly — a 76% satisfaction rate for January–June 2026, ¹³ and recurrence, the metric that actually captures whether the underlying failure was fixed, published only episodically. On the state tier the pendency problem remains structural: 2.13 lakh cases pending as of end-May, with Maharashtra alone carrying a backlog above 35,000. ⁹ A grievance platform that redresses 2.16 lakh cases a month while the same five-lakh complaints keep regenerating is not failing at disposal; it is succeeding at disposal while the failures upstream remain untouched. Publishing recurrence per ministry, per month — the data already exists inside DARPG’s own analysis — would change what ministries optimise for.

The cross-layer angle: The recurring-grievance list is a map of the whole stack’s failure points — Layer 2 (banking fraud resolution), Layer 5 (PM-KISAN), Layer 1 (Aadhaar updates via CSCs). Layer 6 is the feedback loop for every other layer; a feedback loop that measures closure speed but not relief recurrence lets the stack drift while its dashboard stays green.

3. The North-East and Ladakh Get Their Own Governance-Stack Workshop

On September 18, DARPG organised a workshop on Special e-Governance Projects for the North-Eastern states (excluding Assam) and Ladakh at the Central Secretariat’s training arm in New Delhi, under its State Collaboration Initiative — with Director (IT) Sreejith N.P. presenting the ServicePlus and NIC delivery rails that underpin state service platforms. ⁵ The workshop is the latest node in a deliberate strategy: DARPG’s SCI has assisted technology-enabled service delivery, project monitoring, consumer grievance redressal and public safety projects in Meghalaya, Mizoram and Nagaland, showcased as best practices at national e-governance conferences, ¹⁴ and it follows July’s National Conference on NextGen Administrative & e-Gov Reforms in Shillong, which DARPG co-hosted with Meghalaya. ¹⁵

The design logic is in the parenthetical: “excluding Assam” — because Assam’s Sewa Setu is the model the initiative exists to replicate. ⁵ This is capacity building as DPI distribution, and it targets the layer’s real last-mile constraint. The centre can build CPGRAMS, ServicePlus and NeSDA benchmarks; none of it matters without grievance redressal officers and service-delivery backends inside state administrations. That the GRO network has grown from 10,232 officers in 2014 to more than 1.11 lakh by 2025 ¹³ is the quiet numbers story underneath every CPGRAMS headline — each new officer is a door where a complaint can actually land. The consumer test for SCI projects is the same as for any state e-gov platform: not launch events but live service-level data — what share of applications and grievances on, say, a Mizoram or Nagaland platform resolve within the same 21-day norm the Centre holds itself to. The initiative deserves credit for picking the states where that data is hardest to produce.

The cross-layer angle: SCI is Layer 6 distributing Layer 3’s rails — document and service-delivery platforms — to states with thin administrative capacity, with grievance redressal designed in from the start rather than bolted on.

4. Sovereign AI Moves Into the Grievance Interface

The week’s most consequential shift was quiet: a Ministry of Minority Affairs workshop on September 17 titled “AI-Enabled Governance through Indus,” run with Sarvam AI. ⁶ It sits inside a broader ministry AI rollout with Sarvam and Gnani.ai — a 22-language chatbot named Aisha for the ministry’s website, voice AI for the Hajj platform, and AI integration for the PM Vikas scheme. ¹⁶ What makes this a Layer 6 story rather than a procurement note is the platform Indus runs on: Sarvam’s sovereign 105B-parameter model, built under the IndiaAI Mission — the same self-hosted, data-sovereign architecture logic that CPGRAMS applied when it launched Samadhan Didi on May 30, an AI voice chatbot that lets citizens file grievances by speaking in any of 22 scheduled languages, powered by Bhashini and built on government-run infrastructure for data privacy. ^{7 17} NextGen CPGRAMS, now in development, extends the pattern: AI-based categorisation and routing, automated escalation, accessibility features and real-time analytics. ¹⁷

The technology genuinely widens the funnel — voice-first filing removes the literacy and form-navigation barriers that kept the offline citizen out of the grievance system entirely. The risks have moved to the back end. AI triage plus disposal dashboards is exactly the combination that can industrialise the closure-without-relief problem documented in the recurring-grievances analysis: a misrouted complaint, auto-closed inside the clock, is a statistic that satisfies the dashboard and abandons the citizen. The government’s own guardrail language is already on record — the Jaipur Declaration, adopted at the National e-Governance Conference, insists that human intelligence must guide AI in governance. ¹⁸ The operational version of that principle, for this layer, is publishable: AI-triage misrouting rates and human-review sampling results should sit on the same dashboard as disposal counts, and DPDP compliance for grievance data — among the most sensitive personal data government holds — should be audited, not asserted.

The cross-layer angle: Bhashini — built as language infrastructure (Layer 3’s linguistic rails) — is now the front end of Layer 6, while sovereign AI procurement under the IndiaAI Mission ties the grievance layer to industrial policy, and DPDP (Layer 7) now governs the most sensitive data trail in the stack.

The Week in Numbers

• 2,16,827 — grievances redressed by Central ministries/departments in July 2026, per the 51st CPGRAMS monthly report ³
• 13,557 — grievances disposed across Central ministries and States/UTs on September 17 alone ⁴
• 13 days — average Central disposal time in 2026 (January 1–July 15), against the 21-day norm ¹⁰
• 15.2 lakh+ — grievances received on CPGRAMS from January 1 to July 15, 2026 ¹⁰
• 76% — citizen satisfaction rate on CPGRAMS, January–June 2026 ¹³
• 7.36 lakh — public grievances disposed under Special Campaign 5.0 (data as on 11.11.2025) ¹⁰
• ₹816.75 crore — scrap-disposal revenue earned under Special Campaign 5.0 ¹⁰
• 5 lakh+ — recurring grievances across four years flagged in DARPG’s own systemic-gap analysis ⁸
• 22 — scheduled languages supported by Samadhan Didi via Bhashini ⁷