Indian E-Commerce Firm Meesho Expands Vernacular AI Voice Agents Ahead of IPO

Indian e-commerce firm Meesho has developed a real-time voice agent using ElevenLabs text-to-speech to automate customer support in Hindi and English.

It efficiently manages many inquiries without human intervention, covering order delays, cancellations, and refunds, while reducing average handling time, according to Eleven Labs.

“Voice AI is changing how businesses in India interact with their customers. Meesho’s deployment shows what’s possible when multilingual, natural-sounding speech becomes the interface,” said Siddharth Srinivasan, GTM India, ElevenLabs.

The company is set to open subscriptions for its three-day initial public offering (IPO) sale starting on December 3, with a price band of ₹105 to ₹111 per equity share.

Ahead of the IPO, the company’s report said it is deepening its AI capabilities through its Meesho AI Labs initiative, aiming to institutionalise long-term, frontier AI innovation.

The team has built AI-driven systems that personalise the shopping experience, such as optimising product rankings and tailoring recommendations.

Meesho AI Labs has also worked to improve the customer experience through vernacular voicebots that leverage generative AI to resolve queries more naturally and intuitively for shoppers in smaller towns and rural areas.

Key focuses of Meesho AI Labs also include creating language models tailored for Indian consumers, enhancing the shopping journey with agentic AI, improving transaction risk management, and increasing the effectiveness of advertising through AI.

The company intends to build and roll out technology products, including various logistics management systems and content creator technologies, to optimise fulfilment, content generation, and seller operations at scale, Meesho said in a report.

“We are also currently piloting the externalisation of our AI-enabled support services to third parties,” the company added.

Additionally, Meesho Networks was set up in April 2025, “with the objective of commercialising AI technologies and solutions.”

The company intends to continue building its technology stack to “reduce manual overheads and improve process efficiency across the platform.”

They aim to develop technology products, including logistics management systems and content creator tools, to enhance fulfilment, content generation, and seller operations.

Designed for Meesho’s low-cost environment, these systems will leverage automation to expedite cataloguing and product discovery, the report said.

As per Reuters, Meesho aims for a valuation of up to $5.6 billion in its upcoming IPO, which is projected to reach $170-$190 billion by 2030, according to a Bain and Flipkart report.

The post Indian E-Commerce Firm Meesho Expands Vernacular AI Voice Agents Ahead of IPO appeared first on Analytics India Magazine.

Accenture Partners with OpenAI to Roll Out ChatGPT Enterprise for Its Workforce

Accenture and OpenAI have formed a wide collaboration that aims to push enterprise AI deeper into large companies, starting with Accenture’s own workforce.

The firm will roll out ChatGPT Enterprise to tens of thousands of its employees, which the company says is the largest group to be upskilled through OpenAI Certifications. Accenture will use the tool across consulting, operations and delivery work as it builds new AI services for clients.

Accenture has named OpenAI as one of its primary partners for its next wave of AI-powered offerings. The company says the partnership will help it bring agentic AI into the core of business operations and speed up adoption for large enterprises.

Julie Sweet, chair and CEO of Accenture, said, “By combining OpenAI’s breakthrough technologies with Accenture’s deep industry and functional expertise and global delivery capabilities, we will accelerate enterprise reinvention and business outcomes for our clients.”

Fidji Simo, CEO of Applications at OpenAI, said, “Accenture plays an important role in helping companies adopt the technologies that define each new era, and we’re excited to partner with them to accelerate the AI transformation of the largest enterprises.”

The companies are launching a flagship AI program that blends OpenAI’s enterprise products with Accenture’s AI experience. Accenture will get access to OpenAI implementation playbooks, industry use-cases, deployment guidance and hands-on expertise so it can bring AI into real workflows for clients.

Both firms will also work on new solutions for customer service, supply chain, finance, HR and other key business functions. Accenture will use OpenAI’s AgentKit to help clients design and deploy custom AI agents that automate work and support decision making.

According to reports, Accenture has also started rebranding its 800,000 staff and calling them ‘reinventors’ to adapt with AI.

The companies say the goal is to help joint clients adopt OpenAI’s agentic capabilities faster and integrate them across their operations. OpenAI already works with some of the world’s largest enterprises, including Walmart, Salesforce, PayPal, Intuit, Target, Thermo Fisher, BNY, Morgan Stanley and BBVA.

The post Accenture Partners with OpenAI to Roll Out ChatGPT Enterprise for Its Workforce appeared first on Analytics India Magazine.

Legal Heat on TCS, Wipro Prompts Wider IP Rethink in Indian IT

Indian IT’s confidence in its delivery discipline is being tested as TCS faces a major setback in a trade secrets lawsuit, and a fresh patent suit in the US, while Wipro has been named in a new patent-infringement case overseas.

The developments are prompting the industry to reassess whether its governance, engineering culture and compliance frameworks can keep pace with the far stricter intellectual-property expectations of global clients, especially as firms move deeper into platforms, products and proprietary AI systems where IP boundaries are tighter and scrutiny is sharper.

Technology leaders, engineering heads and legal experts say the sector has matured, but gaps remain, particularly, as companies move deeper into platform and product development, where IP boundaries are far tighter than in traditional services.

Ritesh Gupta, CTO of Happiest Minds Technologies, said Indian IT providers have grown significantly more mature in safeguarding client IP as they build products and platforms, driven by stricter client expectations, tighter contracts and heightened IP sensitivity in global technology deals. Happiest Minds Technologies defines itself as “a mindful IT company,” that enables digital transformation for enterprises and technology providers.

According to Gupta, tier-1 firms already operate with strong contractual frameworks, secure development processes and clear IP-ownership governance, and mid-sized firms are rapidly catching up.

At Happiest Minds, he said the company follows rigorous legal vetting of client contracts, secured and isolated development environments, role-based access, regular audits and strict mechanisms for open-source scanning and SBOM management.

Gupta said that engineers receive specific training on IP protection, and that the firm restricts public generative AI tools, while maintaining meticulous code tracing to avoid contamination.

He acknowledged that the industry still struggles with challenges such as unintentional code reuse, weak open-source license governance and risks introduced by generative AI. The services mindset is evolving with the discipline required for product development, he noted.

Gupta said the company’s approach embeds compliance directly into engineering processes, so speed never compromises client IP. He argued that the entire industry needs stronger governance, automated compliance tooling, better awareness and a more IP-conscious culture to prevent violations.

Debashis Singh, chief information officer at Persistent Systems, said the industry’s shift from traditional IT to modern platform engineering has demanded a new level of rigour, as legacy architectures built for stability offer little flexibility and often carry institutional knowledge known only to a few.

He observed that the cloud era introduced open, modular architectures and automation that radically compressed delivery timelines, placing modern engineering at the intersection of speed and discipline.

Singh said Persistent maintains strict access limits and clear IP boundaries through zero-trust frameworks, secure private environments and enterprise-grade AI models that ensure no customer environment is used to train AI systems.

He acknowledged that teams historically struggled to understand undocumented legacy systems, but argued that generative AI and intelligent platforms, such as Persistent’s SASVA, now accelerate reverse engineering and documentation, transforming what was once a slow, opaque process.

According to him, 70–80% of legacy understanding can now be automated, though the final 20–30% still requires human expertise.

He cautioned that reliance on reverse engineering introduces risks such as misinterpreted logic or inadvertent IP exposure, which Persistent mitigates through structured governance, DevOps-led automation and strong knowledge management.

Singh said adherence to IP rules ultimately depends on user awareness, ethical AI guardrails, continuous improvement and embedding compliance into everyday engineering, enabling delivery at speed without compromising legal boundaries.

Priyanka S Kulkarni, a senior legal advisor specialising in IP, said the Indian IT industry broadly recognises the legal risk of exceeding authorised access, especially under US and international regulations. But, it must remain vigilant in documenting permissions, monitoring access and ensuring employees understand the exact boundaries of their authorisations, she added.

She noted that while most vendor–client contracts contain confidentiality clauses, many lack precise definitions of permissible access during modernisation or integration work, creating potential exposure around trade secrets and IP ownership.

According to her, blind spots commonly include overly broad internal access, unclear ownership of accelerators and derivative works, weak exit protocols, and scenarios where multiple vendors share overlapping environments.

She said firms should rely on strict role-based access controls, immutable logs, automated alerts, periodic audits and clear revocation processes to create defensible evidence of compliance.

As companies shift from services to platforms and productised offerings, Kulkarni argued that they must adopt more structured IP frameworks covering background and foreground IP, residual rights, co-developed assets and licensing boundaries, backed by robust data-governance, liability and indemnity clauses. She said the platform era demands far tighter legal, contractual and compliance scaffolding than the traditional services model.

The post Legal Heat on TCS, Wipro Prompts Wider IP Rethink in Indian IT appeared first on Analytics India Magazine.

TwelveLabs Launches Marengo 3.0 Video Understanding Model on TwelveLabs and Amazon Bedrock

Fujitsu Develops Multi-AI Agent Collaboration Tech to Optimize Supply Chains, Launches Joint Trials

KAWASAKI, Japan, Dec. 1, 2025 — Fujitsu Limited today announced the development of a multi-AI…

LTTS Unveils NVIDIA-Powered Digital Twin Platform for Advanced Respiratory Diagnostics

L&T Technology Services (LTTS) has announced the development of a next-generation AI-powered digital twin platform for respiratory diagnostics and lung navigation, developed in collaboration with NVIDIA.

The company said the platform aims to deliver scalable, low-latency solutions that improve diagnostic precision and expand access to advanced imaging tools for clinicians worldwide.

The solution, to be showcased at the Radiological Society of North America (RSNA) 2025 conference, integrates with CT imaging and uses deep learning models to generate a 3D digital twin of lung anatomy.

Built on NVIDIA MONAI for medical image segmentation and NVIDIA TensorRT for optimised AI inference, the platform enables detailed visualisation of airways, blood vessels, lobes and lesions, along with interactive simulation and path-planning support for bronchoscopy procedures.

LTTS said its engineering capabilities in medical imaging and proprietary navigation systems allow static CT scans to be converted into dynamic, clinically meaningful digital representations that evolve with patient data.

The company said such biological digital twins can support planning and navigation for conditions including lung cancer, COPD and infectious diseases.

“AI is reshaping what’s possible in diagnostics and medical technology,” said Alind Saxena, executive director and president of mobility and tech at LTTS.

“Our collaboration with NVIDIA combines LTTS’ expertise in AI-driven diagnostics and predictive analytics, with NVIDIA’s powerful modeling and visualisation platform. This lets us engineer a digital twin platform that not only enhances diagnostic accuracy but also gives clinicians an immersive, real-time planning tool ultimately helping deliver better outcomes for patients around the world,” Saxena added.

David Niewolny, director of business development for healthcare/medical at NVIDIA, said the partnership demonstrates the potential for accelerated computing in clinical workflows.

“Working with LTTS to accelerate the development of AI-enabled medical technology demonstrates how NVIDIA is empowering the healthcare industry with accelerated computing and AI innovation,” Niewolny said.

“LTTS is enabling transformative solutions that bring the vision of real-time AI and biological digital twins powered by NVIDIA to clinical practice, delivering interactive, real-time visualisation and intelligent guidance to help clinicians provide higher-quality care and achieve better outcomes for patients.”

The post LTTS Unveils NVIDIA-Powered Digital Twin Platform for Advanced Respiratory Diagnostics appeared first on Analytics India Magazine.

DeepSeek Releases New Reasoning Models to Match GPT-5, Rival Gemini 3 Pro

Chinese AI lab DeepSeek has launched two new reasoning-first AI models, DeepSeek-V3.2 and DeepSeek-V3.2-Speciale, expanding its suite of systems for agents, tool-use and complex inference.

Both the models and the accompanying technical report have been released as open source on Hugging Face.

The company announced on X that V3.2 is the official successor to V3.2-Exp and is now available across its app, web interface and API. The Speciale variant is offered only through a temporary API endpoint until December 15, 2025.

DeepSeek said V3.2 aims to balance inference efficiency with long-context performance, calling it “your daily driver at GPT-5 level performance.”

The V3.2-Speciale model, positioned for high-end reasoning tasks, “rivals Gemini-3.0-Pro,” the company said. According to DeepSeek, Speciale delivers gold-level (expert human proficiency) results across competitive benchmarks such as the IMO, CMO and ICPC World Finals.

The models introduce an expansion of DeepSeek’s agent-training approach, supported by a new synthetic dataset spanning more than 1,800 environments and 85,000 complex instructions. The company stated that V3.2 is its first model to integrate thinking directly into tool use, allowing structured reasoning to operate both within and alongside external tools.

Alongside the release, DeepSeek updated its API, noting that V3.2 maintains the same usage pattern as its predecessor. The Speciale model is priced the same as V3.2 but does not support tool calls. The company also highlighted a new capability in V3.2 described as “Thinking in Tool-Use,” with additional details provided in its developer documentation.

The company recently also released a new open-weight model, DeepSeekMath-V2. The model, as per the AI lab, demonstrates strong theorem-proving capabilities in mathematics and achieved gold-level scores on the International Mathematics Olympiad (IMO) 2025.

The post DeepSeek Releases New Reasoning Models to Match GPT-5, Rival Gemini 3 Pro appeared first on Analytics India Magazine.

OpenAI Takes Equity Stake in Thrive Holdings

OpenAI, the AI giant and the maker of ChatGPT, has entered into a new partnership with Thrive Holdings (Holdings).

As part of the deal, OpenAI will take an equity stake in Holdings, aiming to bring frontier AI into sectors and industries that work on legacy workflows and systems that have barely changed in decades.

“The partnership will bring together a unique, cross-functional team, comprising OpenAI’s leading research and applied AI teams working alongside engineers, operators, and industry experts at Holdings to deeply integrate AI into the businesses that we own and operate,” said Joshua Kushner, founder of Thrive Capital, in a post on X.

The partnership will involve training advanced AI models for specific tasks within Holding’s businesses, “guided by both company-specific data and expert feedback,” added Kushner.

Other financial details of the partnership were not revealed at the time of the announcement.

“OpenAI now has equity and train models for tasks with company-specific data,” said Sheel Mohnot, a general partner at Better Tomorrow Ventures, in a post on X.

Many others also expressed a similar reaction to the partnership, possibly opening up more avenues of training OpenAI’s models.

Holdings is a holding company created by Thrive Capital in 2025, designed to acquire and operate businesses in traditional service sectors such as accounting and IT.

And Thrive Capital, the New York-based venture capital firm founded in 2009, has backed several high-profile technology companies and is also an established investor in OpenAI.

In 2024, the VC firm led a major $6.6 billion funding round in OpenAI, committing roughly $1 billion.

As of the most recent reports around the share sale of OpenAI employee stock, Thrive was among the firms that bought shares — a deal that reportedly helped push OpenAI’s post-money valuation to around $500 billion.

This has raised fresh questions about the circular nature of the arrangement.

Recent reports suggest that OpenAI’s valuation has surged to $750 billion.

The post OpenAI Takes Equity Stake in Thrive Holdings appeared first on Analytics India Magazine.

Infosys Faces ₹13.6 cr GST Penalty 

Infosys, NVIDIA Introduce Generative AI Powered Telco SolutionsInfosys, NVIDIA Introduce Generative AI Powered Telco Solutions

Infosys on Monday said it received a penalty order of ₹13.60 crore from the Office of the Joint Commissioner of Central Goods and Services Tax (CGST).

In a disclosure to SEBI, the company provided a copy of the order alleging that it had not paid GST on the stay of employees. The claimed violations cover the period from FY19 to FY23.

Infosys informed that the penalty will not have any material impact on its financials, operations, or other activities. Further, they noted that the delay in disclosure was due to the time required to verify the order details and determine the appropriate course of action.

This penalty comes against the backdrop of the GST and tax-related probe that the company has faced over the past two years.

In July 2024, the Directorate General of GST Intelligence (DGGI), following a reference from the Karnataka GST Authority, issued a pre–show-cause notice for alleged tax evasion worth ₹32,403 crore. It pertained to non-payment of IGST under the reverse-charge mechanism on services attributed to the company’s overseas branch offices between July 2017 and March 2022.

Subsequently, in August 2024, the Karnataka authority withdrew its notice and transferred the matter to DGGI for central adjudication. The case was eventually closed by DGGI on June 6, 2025, covering the period from FY19 to FY22.

Infosys also reported receiving a show-cause notice regarding alleged ineligible input tax credit refunds amounting to about ₹415 crore in 2025.

The company challenged this notice in the Karnataka High Court, maintaining that no GST demand was due in this matter.
Meanwhile, in a separate development, its Singapore entity was fined SGD 97,035.90 by the Inland Revenue Authority of Singapore earlier this year due to delays in GST payments, a penalty that Infosys described as immaterial.

The post Infosys Faces ₹13.6 cr GST Penalty appeared first on Analytics India Magazine.