Google and Accel India pick 5 startups and none of them are ‘AI wrappers’

Many AI startup ideas are still little more than superficial “wrappers” built on top of existing models. But as AI model makers add more features, investors are wary of startups that could easily become redundant.

Case in point: When reviewing more than 4,000 applications for a joint AI accelerator for Indian startups run by Google and venture firm Accel, “encapsulated” ideas dominated. But none of them were among the five startups in the final group, Accel partner Prayank Swaroop told TechCrunch (pictured above).

The AI-focused Atoms programme, announced in November by Google and Accel, aims to support early-stage startups to build India-related AI products. Startups selected for the final group will receive up to $2 million in funding from Accel and Google’s AI Futures Fund, as well as up to $350,000 in cloud computing and AI credits from Google, the companies said.

Nearly 70% of rejected applications were “wrappers” — startups that layered AI features like chatbots on top of existing software but “were not reimagining new workflows with AI,” Swarup said.

Many of the remaining applications that were rejected, Swarup said, fell into crowded categories such as marketing automation and AI recruitment tools, areas where investors have seen little newness. Startups in those sectors often struggle to differentiate themselves, he said.

Perhaps this is not surprising. This year’s program received nearly four times the applications as Accel’s previous Atoms groups – with many first-time founders.

India’s growing AI ecosystem is still largely focused on enterprise applications, and Swarup said the applications reflect that. About 62% of submissions focused on productivity tools and another 13% on software development and coding, meaning that about three-quarters of the applications were ideas for enterprise software rather than consumer products. (Swaroop was hoping to see more ideas for health care and education.)

Jonathan Silber, co-founder and director of Google’s AI Future Fund, said the five selected startups closely align with areas where Google expects AI to see deeper adoption in the real world.

The program does not require startups to use Google Forms exclusively, Silber said, noting that many companies combine multiple forms depending on their workflow. The goal, he said, is to collect feedback from startups on how Google Models perform in real-world applications.

Insights from those startups can then be sent to Google DeepMind teams to help improve future models, creating what Silber described as a “flywheel” between startup experience and AI development. “If a company is using an alternative model, that means Google has work to do to build the best model on the market,” he told TechCrunch.

The startups selected for this year are:

• K-Dense, which is building an AI “participant scientist” to accelerate research in areas such as life sciences and chemistry;
• Dodge.ai, which develops standalone agents for enterprise resource planning (ERP) systems;
• Persistence Labs, which focuses on voice AI for call center operations;
• Zingroll, which is building a platform for AI-generated movies and shows;
• Level Plane, which applies artificial intelligence to industrial automation in the automotive and aerospace industries.