Moving beyond traditional models, state police and judiciary have embraced frontier thinking, deploying AI as a transformative force.
On December 12, 2025, the Maharashtra government and Microsoft announced that the tool would be expanded from Nagpur’s 25 police stations to all 1,100 police stations across the state.
Maharashtra’s MARVEL to MahaCrimeOS AI
The state’s MARVEL, which is Maharashtra’s government-owned AI initiative, was created for integrating advanced technology into policing. The Maharashtra government and Microsoft have also signed an MoU to establish three AI Centres of Excellence. As a part of this initiative, MARVEL has worked to adapt CrimeOS AI, developed by CyberEye, a Microsoft AI partner, to finetune it for compliance with state investigation protocols and further configure it in Marathi for easy adoption across the force.
MahaCrimeOS AI ingests complaints in any format, such as PDFs, audio, handwritten notes or images and then uses multimodal intelligence to extract critical information in any language. The system adapts investigation paths, automates analysis and profiles persons of interest with unparalleled speed and efficiency.
This was successfully deployed as a pilot across 23 police stations in Nagpur Rural, including the cybercrime police stations (CCPS). Encouraged by these results, the Maharashtra government now plans to scale MahaCrimeOS AI across the entire state, targeting deployment in all 1,100 district police units.
MARVEL is a special-purpose vehicle (SPV) government entity created in 2024 to introduce AI-based solutions in policing. Chief Minister Devendra Fadnavis had said that the aim of MARVEL was to work with global technology leaders to develop AI Copilots that fundamentally transform governance.
MARVEL was set up to strengthen police intelligence, improve crime prediction and modernise investigation methods, making Maharashtra the first state in India to create an independent AI body for law enforcement.
How does MahaCrimeOS AI work?
Speaking to The Indian Express, Harssh Poddar, Superintendent of Police (Nagpur Rural) and ex-officio director and CEO of MARVEL said, “The platform takes over several routine administrative tasks, which helps officers focus more on solving crimes and assisting victims.”
When an investigator logs in, they see a clear dashboard of all their active cases, any new updates, and pending actions.”
In Nagpur Rural, 127 officers have already been trained to use the system, and it has been used in 261 cases so far. In one notable case, a bank clerk from Saoner in Nagpur district was cheated of Rs 38 lakh. Using MahaCrimeOS AI, officers analysed telecom data and collected open-source intelligence to identify the accused, who were eventually traced to Nashik and a city in Gujarat and were arrested.
One of the most important features of MahaCrimeOS AI is its ability to read a Marathi FIR and automatically generate a detailed investigation plan. The system prepares this plan using Maharashtra Police protocols and guidelines issued by the High Court and Supreme Court for each type of crime. CyberEye (Microsoft Partner ISV and works closely with the Microsoft India Development Center) worked with MARVEL to incorporate the Maharashtra police’s investigation protocols and make the platform available in Marathi, for ease of use by investigators. They then added AI, in the form of an investigation copilot that analyses the case information, generates an automated workflow and suggests investigation pathways.
Poddar explained that earlier, in complex cases such as narcotics, cybercrime, crimes against women or financial fraud, investigating officers often had to wait for senior officials to review files and provide instructions. With the AI copilot, an investigation plan is generated immediately, guiding officers on the next steps, which statements to record, which bank accounts to freeze and what social media profiles to examine.
The platform also analyses telecom data and identifies interlinkages within it, which is crucial for complex crimes. It can gather open-source intelligence for undetected cases by tracing phone numbers across social media platforms and e-commerce websites, helping officers find more information about suspects. It is also capable of preparing time-stamped, date-wise case diaries for submission in court, and it incorporates best practices from judicial guidelines into investigation plans.
Ease of use for officers
The Nagpur Rural team demonstrated the tool to Satya Nadella and Chief Minister Devendra Fadnavis. Officers, including Head Constable Sangita Gawande and API Ashish Thakur, showcased the simplicity of the system.
“The platform is designed primarily in Marathi so that even constables can operate it without difficulty. Routine tasks, such as drafting letters for bank and telecom details, previously took 30 to 45 minutes and required officers to type everything manually. The AI system now generates these letters within seconds and can even email them directly, significantly reducing investigation time,” Poddar said.
How it helps complainants
By automating time-consuming tasks and speeding up investigation planning, MahaCrimeOS AI shortens the overall investigation period. Earlier, advanced open-source intelligence work could only be performed by a few trained officers in the cyber branch.
“With the AI system, all 3,000 officers in Nagpur Rural can now perform these tasks, which means investigations no longer depend on specialists or require officers to travel to headquarters in Nagpur city. This has made the process faster, smoother and more consistent,” Poddar noted.
Legal alignment with new criminal laws
The new Bharatiya Nyaya Sanhita (BNS) requires charge sheets to be filed within 60 to 90 days, depending on the severity of the crime. Since investigating officers often handle 10 to 15 cases at a time, AI tools are essential for meeting these timelines. According to Poddar, MahaCrimeOS AI supports the rapid and structured investigative process required under the new laws and helps officers keep pace with legal changes.
How many cases have been solved?
The AI system has been used in 261 cases so far. In one case where multiple people were defrauded on the pretext of setting up Jio towers on their land, officers were only getting a vague ‘C-type’ location through CDR from call data records. Using open-source intelligence through the platform, they were able to identify the accused in West Bengal and the accused was arrested. (A CDR is a log created by telecom companies every time a phone call, message or data session happens. CDR location means the approximate location of the phone at the time of the call, based on the cell tower it connected to. C-type location means the location was very broad or vague, covering a large area and not enough to pinpoint the suspect precisely.)
The system is currently built for four complex crime categories, namely narcotics, cybercrime, crimes against women and financial fraud, but new models are being developed to expand its use to other kinds of crimes as well.
The Copilot was built using Azure OpenAI Service in Microsoft Foundry and secured with Defender for Cloud. The platform also has built-in access to India’s criminal laws and open-source intelligence, which can be used to link crimes and locate suspects. A key advantage of the platform is its ability to adapt existing best-practice protocols, add a level of knowledge specific to a crime and then develop tailored investigation plans.
Maharashtra chief minister Devendra Fadnavis (left) and Microsoft Chairman and CEO Satya Nadella (right). (Express photo) 
Share
