evidoria

← Back to browse

Good practice Imported

Adalat AI — Kerala's Mandatory AI Transcription Tool for Witness Depositions

India · Kochi · See the India profile

Evidence: Descriptive / self-reported Top 24% 73/100 · Ask Evidence Copilot about this practice

Kerala High Court made Adalat AI, a Malayalam-capable speech-to-text tool, mandatory for recording witness depositions in every trial court from November 2025, after a nine-month, nine-state trial and a four-court Ernakulam pilot — the first such statewide mandate in India.

4 trial courts
Ernakulam pilot courts (from 1 Feb 2025)
9 states
States where tool was tested before Kerala rollout (nine months before Kerala mandate)
Adalat AI — Kerala's Mandatory AI Transcription Tool for Witness Depositions Adalat AI — Kerala's Mandatory AI Transcription Tool for Witness Depositions

Details

Maturity
Scaling
Promoter
Kerala High Court (Kerala Judiciary), with the IndiaAI Mission (Ministry of Electronics and Information Technology); underlying tool built by Adalat AI
Period
Piloted from 1 February 2025 in four Ernakulam courts; mandated statewide from 1 November 2025
Keywords
judiciary, court administration, speech-to-text, e-courts

Context

Manual dictation and typing of witness depositions has long been a bottleneck in Indian trial courts.

Objectives

Adalat AI provides speech-to-text transcription with a specialised Malayalam legal-speech model to record witness depositions, developed with judges, court staff and stenographers and backed by the Government of India's IndiaAI Mission.

Activities

Kerala High Court piloted the tool from 1 February 2025 in four trial courts in Ernakulam district, including sensitive cases involving sexual violence and atrocities. In operation, the tool generates a real-time Malayalam transcript; the witness reviews and signs off before the presiding judicial officer endorses it, and depositions are uploaded to the District Court Case Management System (DCMS).

Results

Under a High Court memorandum issued 27 September 2025, the tool became mandatory in all subordinate/trial courts statewide from 1 November 2025, making Kerala the first Indian state to mandate an AI-driven transcription tool across all trial courts. The tool had reportedly been tested for nine months across nine Indian states before the Kerala rollout.

Conclusions

The public record documents adoption breadth — a nine-state trial followed by a statewide mandate — rather than an independently measured gain in speed, cost, or transcription quality.

Implementation
Indicative cost
High (€500k–€5M) — No public budget or licensing cost figures disclosed; the tool is externally built and vendor-supported (Adalat AI), with backing from the Government of India's IndiaAI Mission.
Time to results
Medium (1–3 years) — Piloted from 1 February 2025 in four Ernakulam courts; mandated statewide from 1 November 2025, preceded by a reported nine-month, nine-state trial.
Staffing & skills
judges, court staff and stenographers involved in tool development, vendor (Adalat AI) providing technical support, state e-Courts cell, presiding judicial officers for endorsement

Conditions for success

  • witness review and sign-off built into the workflow before judicial endorsement
  • integration with the existing District Court Case Management System (DCMS)
  • phased approach: multi-state trial, then a four-court pilot, then a statewide mandate

Common failure modes

  • no published accuracy, error-rate or cost figures, limiting independent assessment of quality

Where it fits

Governance type
state judiciary mandate backed by a national AI mission
Scale
statewide (Kerala)
Income level
lower-middle-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful