AvenuesAI bets on on-premise AI as firms seek tighter control over data
Fintech firm’s PhroneticAI unit builds small language models for deployment within enterprise systems amid rising privacy concerns.
- Apr 15, 2026,
- Updated Apr 15, 2026 11:32 AM IST
AvenuesAI Limited, a listed fintech company, has begun developing fully on-premise artificial intelligence models through its subsidiary PhroneticAI, as enterprises look to keep sensitive data within their own systems.
The company is building a platform based on small language models (SLMs), designed to be deployed entirely within a client’s infrastructure so that no data leaves the organisation’s servers, according to a report by ANI.
Chairman and Managing Director Vishal Mehta said the move is driven by growing concerns around data security and privacy in the use of AI tools.
“Companies are increasingly concerned about data security and privacy while using artificial intelligence tools,” Mehta told ANI.
The development comes at a time when firms are reassessing their reliance on AI systems that depend on external servers, particularly as questions around data control and sovereignty gain prominence.
“Developments involving OpenAI and Anthropic have raised concerns globally about data independence and control,” Mehta said.
He added that enterprises handling sensitive information are becoming more cautious about how their data is processed. “Companies handling sensitive information are increasingly evaluating whether their data could be exposed when processed through third-party platforms,” he said.
The push for on-premise AI has also been shaped by recent disruptions in outsourced digital services. The report cited a mid-2025 incident involving Nayara Energy, where Microsoft temporarily suspended cloud and email services, citing compliance with European Union sanctions.
Nayara challenged the move in the Delhi High Court, and services were restored within days, but the episode highlighted potential risks in relying on external infrastructure.
Against this backdrop, AvenuesAI is developing SLMs ranging from 1 billion to 10 billion parameters that can be trained on a client’s own data and deployed within internal systems.
The platform is designed to handle enterprise data across formats, including documents, spreadsheets, audio, video and images, while offering tools for monitoring and retraining models within the organisation’s infrastructure.
Mehta said the company’s approach is to enable clients to operate independently once the system is deployed. “The company aims to build systems that clients can operate independently after deployment, without requiring continued backend access from the vendor,” he said.
Despite contractual safeguards, hesitation around external AI providers remains. “Companies remain cautious about sharing sensitive data with external AI providers despite contractual safeguards,” Mehta added.
AvenuesAI’s push into on-premise SLMs reflects a broader shift in enterprise AI adoption, where control, compliance and data localisation are emerging as key decision factors alongside performance and cost.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine
AvenuesAI Limited, a listed fintech company, has begun developing fully on-premise artificial intelligence models through its subsidiary PhroneticAI, as enterprises look to keep sensitive data within their own systems.
The company is building a platform based on small language models (SLMs), designed to be deployed entirely within a client’s infrastructure so that no data leaves the organisation’s servers, according to a report by ANI.
Chairman and Managing Director Vishal Mehta said the move is driven by growing concerns around data security and privacy in the use of AI tools.
“Companies are increasingly concerned about data security and privacy while using artificial intelligence tools,” Mehta told ANI.
The development comes at a time when firms are reassessing their reliance on AI systems that depend on external servers, particularly as questions around data control and sovereignty gain prominence.
“Developments involving OpenAI and Anthropic have raised concerns globally about data independence and control,” Mehta said.
He added that enterprises handling sensitive information are becoming more cautious about how their data is processed. “Companies handling sensitive information are increasingly evaluating whether their data could be exposed when processed through third-party platforms,” he said.
The push for on-premise AI has also been shaped by recent disruptions in outsourced digital services. The report cited a mid-2025 incident involving Nayara Energy, where Microsoft temporarily suspended cloud and email services, citing compliance with European Union sanctions.
Nayara challenged the move in the Delhi High Court, and services were restored within days, but the episode highlighted potential risks in relying on external infrastructure.
Against this backdrop, AvenuesAI is developing SLMs ranging from 1 billion to 10 billion parameters that can be trained on a client’s own data and deployed within internal systems.
The platform is designed to handle enterprise data across formats, including documents, spreadsheets, audio, video and images, while offering tools for monitoring and retraining models within the organisation’s infrastructure.
Mehta said the company’s approach is to enable clients to operate independently once the system is deployed. “The company aims to build systems that clients can operate independently after deployment, without requiring continued backend access from the vendor,” he said.
Despite contractual safeguards, hesitation around external AI providers remains. “Companies remain cautious about sharing sensitive data with external AI providers despite contractual safeguards,” Mehta added.
AvenuesAI’s push into on-premise SLMs reflects a broader shift in enterprise AI adoption, where control, compliance and data localisation are emerging as key decision factors alongside performance and cost.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine