How agentic AI could improve enterprise data operations

Without good data, AI isn’t going to be as useful as it should be to an enterprise. Managing and optimizing data workflow however is not an easy task, but it might be getting a bit easier thanks to the power of AI agents.

San Francisco-based startup Altimate AI today announced its new DataMates technology which brings the concept and power of agentic AI to enterprise data operations. The basic idea behind DataMates is to provide AI agents that help enterprise data teams automate and accelerate a wide range of tasks, from data documentation to performance optimization. The goal is to reduce the burden on overworked and understaffed enterprise data teams so they can optimize data operations to meet business requirements. Data operations are critical not just for AI, but also for ongoing business intelligence, operations and data analytics. The challenge is that there is an ever growing volume of data, but not an ever growing volume of staff.

Altimate AI was founded in 2022 and benefits from the backing of John Chambers, the former CEO of Cisco who currently runs JC2 Ventures, which has invested in the company. Altimate AI already has its DataPilot platform in the market which provides data automation capabilities. The new DataMates service will be an integrated part of DataPilot, accelerating data operations with AI agents.

“Many times we saw data teams bottlenecked with lot of work, because there are always analytics and AI projects that are there in the pipeline, and just the amount of work they need to do is humongous,” Pradnesh Patil (CEO) and co-founder of Altimate AI told VentureBeat in an exclusive interview.  “We started the company with the vision of accelerating and automating the work that they do.”

Bringing the power of Agentic AI to enterprise data teams

Patil explained that DataMates is intended to act as a virtual teammate for overworked enterprise data teams.

DataMates act as autonomous members of data teams that can perform tasks that are often time consuming and repetitive for humans. Common data operations tasks that DataMates promises to handle automatically include documentation, testing and data transformations. The tasks are not just simple automation scripts either, which isn’t something new for enterprise data engineers. Rather Patil emphasized that DataMates uses agentic AI to pull lots of context from the company’s entire data stack in order to execute any data related task. He noted that the contextual understanding allows Data Mates to perform tasks with a level of nuance typically associated with human experts.

How DataMates works to save enterprise data teams time

The DataMates system is built on a proprietary framework that combines multiple language models, function calling mechanisms and a custom-built knowledge graph.

“We basically have our own in-house framework for a bunch of things,” Anand Gupta, CTO and co-founder of Altimate AI told VentureBeat. 

Gupta said that Altimate AI does also make use of commerical LLMs, though he did not specifically identify which models were being used. He did however emphasize that the agentic AI framework that pulls together multiple LLMs for data operations is something the company mostly built in-house.

This architecture enables data mates to perform a wide range of tasks, including:

Data model drafting

Automated documentation

Test generation and execution

Code review and optimization

Performance analysis and tuning

In terms of how enterprise data professionals can access DataMates, the idea is that it easily fits into the normal workflow with existing tools such as VScode, Git and Slack.

“These AI teammates, we are making available right in the tools that data teams use, so that they literally act as teammates sitting next to you,” Gupta said. “You can basically give some key tasks to them, they do it autonomously and in that way, the workload on your plate is much lesser, and you can accelerate your project delivery.”

Agentic AI is great, but Ambient AI makes it better

Going a step further than just automating tasks that data professionals need done, Altimate AI is also integrating a form of ambient AI.

Patil explained that the ambient AI layer acts as a continuous, intelligent monitoring system that analyzes the data infrastructure and provides proactive suggestions to the data team.

He noted that the goal of the ambient AI is to provide these insights and recommendations without requiring constant manual intervention from the data team. This allows the data team to focus on higher-level tasks while the ambient AI handles the ongoing optimization in the background.

Why the former CEO of Cisco John Chambers is all in on Altimate AI

The promise of Altimate AI has attracted numerous enterprises as well as a noteworthy enterprise investor, with John Chambers, the former CEO of Cisco who currently leads JC2 Ventures.

“My son is the two in JC2,” Chambers told VentureBeat in an exclusive interview. “He worked with one of the two founders, and he said, Dad, this guy’s as smart as it gets.”

Chambers’ son had previously worked at Walmart Labs, which is where he met Gupta. The problem Altimate AI is solving was one that Chamber’s son had experienced at Walmart Labs, which gave him insight into the value of the solution. The problem was about getting data into the right format quickly in order to run marketing programs.

In Chambers’ view, Altimate AI has a differentiated solution in a high-growth market. He is particularly interested in Altimate AI’s focus on improving data engineer and data scientist productivity, which he sees as a rapidly growing and underserved market segment. 

“There are only a few startups really focusing on this, and while the big players are dabbling in it, so far, they haven’t moved, so Altimate AI has a first mover advantage,” Chambers said. “And it’s being done by people who had this problem, not who theoretically see it.”