प्रगति साथी — NIPUN Bharat Impact Pulse Dashboard (Demo)

Surveys completed412this pulse round (synthetic)

Response rate63%of connected calls finished

Training adoption78%teachers: NISHTHA FLN done

TLM in use54%teachers: available + used

Home reading ≥3 d/wk41%parents: child reads aloud

Mean satisfaction3.8/5teachers + parents

Live demo — सर्वे कॉल का अनुभव लें

Click the orb and role-play a respondent. प्रगति साथी runs the neutral survey; after the call, ElevenLabs extracts the structured fields below from the transcript automatically — that extraction is the KPI pipeline.

Role-play one of these

The agent branches by respondent type — try both tracks.

शिक्षक: "मैं शिक्षक हूँ, सीतापुर से।" — फिर ट्रेनिंग, खेल-गतिविधि, TLM, प्रगति, रुकावट और संतुष्टि के सवालों के जवाब दें।

अभिभावक: "मैं अभिभावक हूँ, लखनऊ से।" — फिर निपुण जागरूकता, घर पर पढ़ना, PTM और संतुष्टि के जवाब दें।

Extracted per call: respondent_type · district · training_received · methods_adopted · tlm_used · reading_progress · barrier · nipun_aware · home_reading_days · ptm_attended · satisfaction · suggestion

प्रगति साथी से बात करें

Neutral surveyor · ~3 minutes · answers go to the dashboard, not spoken back

Click the orb to start the call

प्रगति साथी identifies itself as an automated AI helper for an education impact study — never as a person or government officer. Demo only; no real respondent data.

Input → practice funnel · शिक्षक ट्रैक

The mission's theory of change, measured: training received → play/activity methods adopted → TLM actually used. The drop-offs show exactly where support is needed (synthetic cohort, n = 236 teachers).

Adoption funnel

% of surveyed teachers

District comparison · TLM usage

% teachers using TLM, by district

Reported barriers

"सबसे बड़ी रुकावट क्या है?"

Trend vs external anchor

Pulse-round KPIs tracked monthly, anchored against ASER's biennial reading measure (Std-III govt-school children reading Std-II text: 16.3% in 2022 → 27.9% in 2024). The pulse survey fills the gap between those rounds.

Monthly pulse trend (synthetic)

Teacher method-adoption % and parent home-reading ≥3 d/wk %

Recent extractions

First two rows are REAL extractions from simulated test calls to this agent; rest synthetic.

Respondent	District	Key signals	Satisfaction
शिक्षक sim	सीतापुर	ट्रेनिंग ✓ · methods ✓ · TLM ✗ (class size)	4/5
अभिभावक sim	लखनऊ	निपुण aware ✗ · reading 2 d/wk · PTM ✓	3/5
शिक्षक	हरदोई	ट्रेनिंग ✓ · methods ✓ · TLM ✓	5/5
अभिभावक	सीतापुर	निपुण aware ✓ · reading 4 d/wk · PTM ✗	4/5
शिक्षक	बाराबंकी	ट्रेनिंग ✗ · TLM ✗ (training gap)	2/5
अभिभावक	उन्नाव	निपुण aware ✗ · reading 1 d/wk · PTM ✗	3/5

How the impact pipeline works

STEP 1

आउटबाउंड कॉल

Agent calls sampled teachers & parents (Vobiz SIP trunk for India outbound — already live in this stack). Neutral 3-minute Hindi survey, one question per turn.

STEP 2

Structured extraction

After each call, the platform extracts typed fields (booleans, numbers, strings) from the transcript — no human data entry.

STEP 3

KPI aggregation

Post-call webhook streams extractions to the dashboard: adoption %, TLM usage, home-reading frequency, satisfaction, barriers — by district and month.

STEP 4

Decision support

Drop-offs in the funnel target follow-up (e.g. TLM gap in a district → BRC action), between NAS/ASER rounds instead of years later.

Demonstration disclosure. All aggregate numbers are synthetic except the two marked simulated-call extractions. Scheme grounding: NIPUN Bharat targets every Grade-3 child attaining FLN by 2026-27 (NCERT mission guidelines); current impact measurement (NAS / Foundational Learning Study / ASER) is sample-based and 2–3 years apart — this system is proposed as a continuous screening-grade pulse, not a replacement for formal assessment. Phone-survey results carry response bias; AI disclosure is spoken in every call. No real respondent data was used.