Mohit Pandey

For ServiceNow, the next enterprise AI battle is not about who has the smartest model. It's about who can keep thousands of AI agents under control. GitHub has finally brought Stacked Pull Requests into the platform, but the launch has sparked a wider conversation about how code review should evolve as AI increasingly writes more code.

Article Complete Digi Yatra CEO Suresh Khadakbhavi isn't really obsessed with airports. His real obsession has always been removing friction, wherever it exists. Why India's Most Advanced Weather Models Still Send Alerts Late Despite better weather models, growing AI capabilities, and national alert systems, Kerala's recent flash floods reveal that India's biggest challenge may not be predicting disasters. AUGUST 8, 2026, 10:00 AM 5 min Key Takeaways What Actually Matters. India's advanced weather models and AI capabilities struggle to deliver timely disaster alerts. Kerala's flood response highlights failures in basic alert systems despite technological advancements. Merin Susan John Associate Technology Journalist Followed by 12 readers Got a tip? Share confidential information with AIM.

Two, by his own admission, belonged to people who clearly did not know what they were talking about. He was not sure about the third either. The centre will house capabilities including software development, data and AI.

MWire Labs has built a multilingual speech AI system for languages including Khasi, Garo, Mizo, Kokborok, Assamese, and Nagamese, using more than 1,000 hours of field audio. The organisation will use the accelerator's backing to expand its speech recognition, case management and paperless courtroom tools across India and other low-resource legal systems.

For now, the company is choosing caution over unrestricted access. Astra is expected to become available soon, but its most powerful cybersecurity capabilities will remain limited initially. Currently, about 25% of iPhones are assembled in India. But it all started when Tim Cook made his first official visit to India in 2016.

Before Article Content · 728×90 Advertise Here Article Bottom · 728×90 Advertise Here

Article Complete YC Backs India's Legal Tech Nonprofit Adalat AI As It Takes Courtroom AI Global Adalat AI says it now works across 11 states and has reached more than 6,000 judges, covering roughly 20–25% of courtrooms nationwide. SEPTEMBER 3, 2026, 4:33 PM 5 min For Adalat AI, the timing is significant. Its Co-Founder and CEO, Saxena, was named among TIME's 100 Most Influential People in AI for 2026 just days after the YC announcement. TIME's recognition puts an unusual spotlight on a category of AI that rarely gets the attention of frontier models, chips or billion-dollar infrastructure projects: technology being built for public institutions. Now, with YC backing the company, Adalat AI wants to take what it has built in Indian courtrooms beyond India. Its founders, Saxena and Bhattacharya, will spend September to December in San Francisco as part of the programme. For Bhattacharya, the CTO, YC is less about changing what Adalat AI is building and more about helping the organisation figure out how to take it to a much larger scale. “Well, you never quite expect any of this, especially given we are a nonprofit, and that's generally not what happens. But we are extremely grateful to YC for knowing this opportunity, and we're really looking forward to learning how to scale this and learning the best of how to build technology from them,” Bhattacharya tells AIM. Why This Problem? Adalat AI has been building AI infrastructure for courts for roughly two and a half years. Its technology includes live speech-to-text transcription, case-flow management and paperless court workflows. The company says it now works across 11 states and has reached more than 6,000 judges, covering roughly 20–25% of courtrooms nationwide. Bhattacharya says that when he first met Saxena, he did not immediately believe that something as simple as speech-to-text could make a meaningful difference. He changed his mind after entering a courtroom. “We were walking past this courtroom, and there were these 10 foot towers of paper inside the rooms. And then there was a person standing outside the room, and their full-time job was to take chits of paper from outside, and then look at the chit of paper, and then spend the next few hours looking for the right files amongst those towers of paper,” Bhattacharya says. Its first product was built around one of the most basic bottlenecks inside courtrooms: transcription. In many courts, there is no stenographer. Bhattacharya says only about 10% of Indian courts have one. That makes courtroom speech recognition much harder than simply putting a voice assistant in front of a judge. A system that works well in a controlled environment can fail quickly inside a district courtroom. “We needed to build a system and move away from a pilot to a system that was actually secure, ready, reliable, and could yield the benefits and the promise that transcription was making in the lives of judges.” But Technology Alone Was Not Enough This is where being a nonprofit became part of Adalat AI's strategy rather than simply its legal structure. Courts are deeply institutional environments, and convincing judges and court staff to adopt a new technology requires a degree of trust that cannot be created by a better model alone. Bhattacharya describes Adalat AI as a “twin-engine system”. One side is technology. The other is its partnerships team, which works directly with judges and court staff, collects feedback and helps them use the systems. That approach appears to be paying off. Kerala's High Court mandated Adalat AI's speech-to-text system for witness depositions in November 2025. Sikkim subsequently became the country's first paperless High Court, with digital filing becoming part of its workflow. Andhra Pradesh has also announced a statewide rollout covering transcription, case management and courtroom workflows from October. No Judgements Adalat AI does not decide cases, predict outcomes or recommend sentences. Even its transcription system is designed around keeping a human in the loop. The concern is obvious. If an AI system incorrectly transcribes a witness deposition, the mistake could potentially affect a case. “There is no world where every line is not checked by the lawyers itself. So even if a model gets something wrong, which it does very little now, our transcription rates have been improving quite a bit. It's always scrutinised. There is someone correcting the output,” Bhattacharya says. The company says its early measurements suggest judges using Adalat AI can become two to three times more productive. Its initial studies also suggest that this could reduce overall case-resolution timelines by around 30%. Those figures are still being tested, and the company is conducting a randomised controlled trial to measure the longer-term impact on productivity, case throughput and undertrial populations. Adalat AI has so far been funded through CSR partnerships, fellowships and prize money rather than the conventional venture-capital route. YC gives it access to a network and a playbook built around scaling technology organisations, even though the organisation itself is structured as a nonprofit. Bhattacharya sees two distinct directions for that scale. The first is India. “We want to do this in all courtrooms in India, and I see a future where you know you can file cases in any state, in any district online,” he says. The second is the Global South. Many countries across Africa and other parts of the developing world inherited legal and administrative systems shaped by colonial rule. “It's almost like a copy paste artifact, you know, the laws as well. And so that's the second dimension along which we aspire to scale, which is you know kind of have this be that make in India product that we build in India and then sort of scale and contribute back to the rest of the world as well.” That ambition also gives the YC announcement a larger significance. Saxena's inclusion in TIME100 AI had already placed Adalat AI in a global conversation about what influential AI work looks like. His experience as a Supreme Court law clerk and lawyer led him towards the administrative machinery behind justice, and then towards AI as a way of improving it. For now, the company remains cautious about moving from administrative assistance into legal reasoning. There may eventually be room for AI to handle some categories of cases or processes that require little judicial discretion, but Bhattacharya described that as an idea rather than a current product direction. “I think we need to bring back a culture of building tech for good and making the idea of building tech for good sexy. And I think there is so much good that tech can do. We don't realise how much diffusion work is left of diffusing all this good tech and AI solution that we're kind of building in the systems that need it the most.” With YC now behind it, Adalat AI has a chance to test that idea at a much larger scale. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought • • • • • Mohit Pandey Journalist Mohit writes about AI in simple, explainable, and often funny words. He's especially passionate about chatting with those building AI for Bharat, with the occasional detour into AGI. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy Bengaluru's Makr Microsystems Raises ₹10.2 Cr Led by Bluehill VC The funding will support Makr's technology development, customer validation and path towards commercial deployment. SEPTEMBER 3, 2026, 4:33 PM 5 min Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought

Article Complete [Exclusive] Govt Eyes ₹20,000 Cr Fund to Back India's Frontier AI Race [Exclusive] Govt Eyes ₹20,000 Cr Fund to Back India's Frontier AI Race The NFAICF is being structured as a SEBI-regulated Category I Alternative Investment Fund. SEPTEMBER 24, 2026, 12:32 PM 5 min The proposed fund aims to provide long-term risk capital to India's frontier AI companies as well as finance GPU clusters, specialised data centres, and other infrastructure required to build and scale these models. The proposal is still under consultation, with the final size, structure, governance, and role of government capital yet to be decided. The proposed fund was discussed in a closed-door meeting on ‘Financing India's Frontier & Compute Ecosystem' in Delhi on September 24. The session was hosted by Mohammed Y Safirulla K, Director, IndiaAI Mission; Sudeep Shrivastava, COO, IndiaAI Mission; and Saurabh Vijay, Chief Executive Officer, IndiaAI Mission. The meeting was attended by several Indian AI startups and AI ecosystem stakeholders, including Vivek Raghavan of Sarvam AI, PeakXV MD Rajan Anandan, and S Krishnan, Secretary, Ministry of Electronics and Information Technology. According to a study by research institute Epoch AI, training large frontier models could cost more than $1 billion per run by 2027. The government says costs at this level could put frontier AI development beyond the reach of existing Indian venture capital and grant-based funding. While the IndiaAI Mission offers access to subsidised compute and funding for indigenous foundation models through its ₹10,372 crore outlay, the government notes that there is currently no dedicated domestic equity vehicle to back frontier AI development. That's where it is trying to position NFAICF. The proposed fund intends to address the financing gap with patient capital over long-horizon research, development, and commercialisation cycles. The NFAICF is being structured as a SEBI-regulated Category I Alternative Investment Fund. The investment decisions will be made by an Investment Committee comprising experts in technology, investment, infrastructure, and finance. The proposed fund could consider different investment instruments, including venture and growth equity, convertible instruments, infrastructure equity, fund-of-funds commitments, and direct co-investments, depending on the company or infrastructure project. The aim of the fund is two-fold. The first is providing long-term capital to foundation model builders, specialised AI systems, and other strategically important AI technologies. The second is financing GPU clusters, specialised data centre capacity, and other AI compute infrastructure, while accounting for rapidly changing hardware, utilisation, power requirements, and demand visibility. The government is considering linking compute investments to actual demand from AI companies and research institutions through capacity commitments, utilisation arrangements, or pre-agreed commercial terms. It estimates that this can reduce the risk of stranded or underused infrastructure. The stakeholders are also examining secondary-market mechanisms that could give investors an exit without forcing companies into premature outcomes, given the decade-long investment horizons in traditional deeptech companies. The government is also considering a follow-on and co-investment capability so the fund can participate in later rounds alongside private investors when strategically important companies need more capital. The proposal is now being tested with frontier AI companies, data centre operators, cloud providers, GPU providers, domestic VCs, private equity funds, pension and insurance investors, multilateral institutions, and policymakers. The final design remains subject to stakeholder feedback and further assessment. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought Our Coverage of AI News • Coremantle co-founder Vivekanand Pani believes the AI industry is looking at the Indian language data problem the wrong... Read more → • According to Adobe's India AI and Digital Trends research, 65% of Indian consumers already use AI for personalised... Read more → • “Even the most commoditised model cannot operate in a vacuum. It has to work on top of an enterprise's own foundation... Read more → About the Author Mohit Pandey Journalist Followed by 25 readers Mohit writes about AI in simple, explainable, and often funny words. He's especially passionate about chatting with those building AI for Bharat, with the occasional detour into AGI. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy OpenAI, Anthropic CEOs Urge UN To Set Global Safeguards For AI OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have called for global AI safeguards, warning that poorly managed AI could pose risks to humanity. SEPTEMBER 24, 2026, 12:28 PM 5 min OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have urged the United Nations to establish safeguards for artificial intelligence, warning that the technology could become difficult to control as its capabilities expand. Speaking at a UN Security Council meeting on AI, Altman warned humanity “could lose control of the future to AI,” while Amodei called poorly managed AI “a risk to humanity as a whole.” Both urged the need for safeguards, including common testing standards and measures to verify advanced AI capabilities. The discussion at the UN comes as governments grapple with how to manage increasingly capable AI systems without slowing their potential benefits. Altman and Amodei both argued that AI should not become concentrated in the hands of a single company or country. Altman also highlighted the potential of AI to improve people's lives, while warning about the risks of losing control over its development. Amodei called for countries to agree on measures to prevent AI from being used to develop biological weapons. He also proposed ways to verify the capabilities of advanced AI models and establish common standards for testing them, including tests for potential misuse and loss of control. “Worries about control of AI are not a reason to pause its further development or to constrain it with new global governance structure,'” said White House Science Adviser Michael Kratsios. “We want you, our allies and partners, to share in the benefits of this technology,” he added. China also weighed in, with UN Ambassador Fu Cong saying countries should be able to choose AI technologies according to their needs. The US and China are also discussing a possible mechanism to notify each other about AI incidents that could affect national security. The UK is pushing for greater coordination, with Foreign Secretary Ed Miliband saying AI will be a major focus when the UK leads the G20 next year, including efforts to establish common global standards for safe AI development. The debate also extended to military applications. Ukrainian President Volodymyr Zelenskyy warned that AI could increasingly influence battlefield decisions and called for human control over machines. Experts also raised concerns about autonomous weapons and accountability when AI is involved in decisions involving the use of force. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought About the Author Kileuna Haralu Contributor Kileuna is a journalist covering the real-world impact of artificial intelligence at the intersection of tech, policy, and society. Got a tip? Share confidential information with AIM.

Article Complete Shodh AI Cut Factory Optimisation From 500 Hours to 5.6 With LUCAN Shodh AI's LUCAN wants to change what physical AI does on the factory floor—but by not simply predicting what a physical system will do. SEPTEMBER 24, 2026, 11:45 AM 5 min The model-generated intervention was first independently checked against classical numerical physics solvers before being taken into a physical industrial environment, according to Shodh AI's technical whitepaper. The company is now taking LUCAN beyond chemical and biological manufacturing, with a pilot planned with semiconductor equipment maker Lam Research. Unlike conventional AI models that predict the outcome of a given process, LUCAN is designed to work backwards from a desired physical outcome and identify the molecular and process variables that should change to achieve it. Shodh AI Founder and CEO Arastu Sharma describes this as cross-scale inverse design. From Simulation to Intervention Industrial engineering today relies heavily on specialised tools. Molecular simulations, thermodynamics, computational fluid dynamics, biological mechanics, and process controls are typically handled separately, with engineers moving information between them and repeatedly running forward simulations. LUCAN attempts to combine these layers. The model uses a Sparse Mixture-of-Experts architecture that can work with different scientific representations, including molecular graphs, three-dimensional physical fields, and deforming geometries. These are mapped into a shared computational system so the model can reason across molecular, process, and equipment scales. The distinction is between asking what will happen if a particular parameter changes and what needs to change to achieve a particular result. As Sharma explains in a conversation with AIM, “I'm able to put the outcome that I want and ask it to tell me what parameters I should go with from the start in all of these worlds.” The whitepaper describes the conventional engineering workflow as “propose → simulate → evaluate → adjust → repeat”. LUCAN is designed around “specify outcome → generate intervention → verify.” That means a manufacturer could specify a target yield, for example, and have the model identify the process conditions that could move the system towards that target. The Aarti Industries Test Shodh AI initially focused on chemical and biological manufacturing because India already has significant industrial capacity in these areas. One of its early industrial partners was Aarti Industries, which manufactures speciality chemicals. The company was looking to move from traditional batch manufacturing towards continuous-flow processes, where reaching the desired yield can require years of engineering and experimentation. It used LUCAN to work backwards from the desired industrial outcome, resulting in increased yield and downsizing optimisation process. The validation process is central to Shodh AI's claim. The company does not position LUCAN as a replacement for established physics simulators. Instead, it uses those systems as an independent benchmark before an intervention reaches the physical world. “These simulators exist for a reason. They are calculators. They're amazing at their job,” Sharma asserts. The whitepaper lays out three levels of evidence: a frozen model tested on predefined computational tasks; a model-generated intervention independently checked against high-fidelity classical physics; and a model-generated intervention physically executed in a laboratory, pilot, or industrial environment. The Aarti result falls into the third category. For industrial customers, raw AI benchmark performance is not the primary concern. Companies want to know whether a model agrees with the physics they already trust and, ultimately, whether its recommendations work in the real world. That is why Shodh AI deliberately compares LUCAN against established numerical solvers rather than other AI models. The model can potentially produce a prediction much faster than a traditional simulation while also enabling inverse design and calibration with enterprise-specific data. Why Sales Needs Physical AI A major challenge in industrial manufacturing is that physics changes with scale. A reaction that works in a test tube may behave differently in a larger reactor because fluid dynamics, heat transfer, mass transfer, and other physical effects change. LUCAN is designed to connect these scales rather than model them independently. The whitepaper says information can propagate upward from microscopic energetics to reactor-scale thermodynamics, while information can also move downward from reactor-level conditions towards microscopic structural responses. This is important for inverse design because thousands of variables can influence an industrial process. “If an AI model is able to reason across all of these parameters at the same time—from the quantum parameters to the larger parameters—then it's not limited to a human figuring out the scaling problem,” Sharma remarks. Shodh AI says its current model has grown from 2 billion parameters to 5 billion parameters within its IndiaAI Mission project. Sharma said the company is currently building a model in the 10–15 billion parameter range and wants to scale it towards 100 billion parameters. The Data Security Problem Shodh AI is also trying to address the data-security problem in deploying AI inside regulated industries. Rather than asking customers to hand over large datasets to train the foundation model, LUCAN is trained using physics-based simulation systems and can then be deployed within an enterprise environment and calibrated with relatively small amounts of customer-specific data. “If, say our model is 98% accurate to the physics of the world or that factory, it just needs that 2% calibration that can be done within the enterprise itself,” the founder explains. The goal is eventually to deploy these systems inside industrial environments while keeping sensitive data within the customer's infrastructure. From Chemicals to Semiconductors Shodh AI's next major target is semiconductor manufacturing. The Indian government recently connected the startup to Lam Research, which has agreed to conduct a pilot. Lam Research makes semiconductor manufacturing equipment, but customers still need to spend significant amounts of time optimising those machines for their own manufacturing processes. That is where Shodh AI sees an opportunity for LUCAN. India's semiconductor ambitions make the problem particularly relevant. Building manufacturing capacity requires not only acquiring equipment but also developing the processes needed to operate it efficiently. Sharma estimates that a semiconductor prototype at the relevant scale can cost around ₹30 lakh per attempt, making physical trial and error expensive. The company believes physical AI could reduce both the cost and time involved in developing those processes. The Bigger IndiaAI Bet Shodh AI was selected under the IndiaAI Mission's foundation-model programme, and Sharma says access to compute remains one of the biggest requirements for scaling LUCAN. The company is discussing access to around 2,000 GPUs for a year. But Sharma argues that compute is only one part of what India needs to build physical AI. He wants faster government decision-making, more willingness to take risks on deep technology, and greater access to public-sector and industrial environments where startups can test their models. Shodh AI earlier worked with Indian Oil on a paid pilot and is exploring additional industrial deployments. LUCAN ultimately wants to change what physical AI does on the factory floor—by not simply predicting what a physical system will do, but figuring out what humans should change to make it do something different. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought • Mohit Pandey Journalist Mohit writes about AI in simple, explainable, and often funny words. He's especially passionate about chatting with those building AI for Bharat, with the occasional detour into AGI. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy US Biopharma Co Syneos Health Opens 137,000 Sq Ft Global Life Sciences and Technology Centre in Hyderabad The centre is Syneos Health's largest facility in Asia and third-largest globally. SEPTEMBER 24, 2026, 11:24 AM 5 min Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought

ServiceNow CTO Pat Casey believes the next phase of enterprise AI will be won by making AI perform work that already exists inside a company.