Software developer Turing Labs’ new suite of domain trained AI agents launching in August promise to jumpstart stalled innovation in CPG by taking companies beyond product conceptualization to deliver “execution excellence,” according to company CEO Manmit Shrimali.
Post-pandemic, CPG companies, retailers and co-manufacturers have struggled to deliver meaningful innovation for myriad reasons, including supply chain and labor challenges, constricted capital and years of disinterest by distributors that focused on the most popular SKUs during the height of COVID-19 when consumers were desperate for basics.
Not included in the long list of factors contributing to the innovation gap is a lack of ideas, which AI and machine learning can generate by the dozens seemingly in the blink of an eye. Rather, the main problem that continues to hold back innovation, is successful execution, according to Shrimali.
“Recently there was a customer that used AI to generate a lot of concepts – 15 concepts – but what eventually happened is that only 10 were meaningful and none of them actually were launched,” he explained, noting this is not an isolated experience.
He attributes this lack of “execution excellence” in part to CPG companies’ over reliance on electronic workbooks and “the promise that you put all your data into a magic box and one day you will be able to use this for AI.”
That day has yet to come, he said, and now companies are saddled with legacy systems that “everybody hates, but for which there has been no other option.”
He added, this is compounded by rising pressure for full “digital transformation,” – a daunting task for which few companies know where to begin or how to efficiently and effectively overhaul their complete systems.
Enter Turing’s Luna AI – an “intelligent R&D assistant,” which the company is upgrading with the creation of “the industry’s first domain trained agents that can be configured and deployed at scale for all of these enterprises,” according to Shrimali.
He explained Luna AI “starts with the KPI of the company” and pre-configured and orchestrated agents systematically tackle each task to reach commercialization “in a very systematic way.”
Not only are the agents “highly trained” for innovation,” but “they can be configured for digital transformation” so companies do not have to think simultaneously about what data they need to collect – “these agents are purposely built to take care of all those kinds of things.”
What does Luna AI look like in action?
Using Luna AI’s new R&D agents is as easy as typing a question or KPI into a search field, according to Shrimali.
“The way we are helping first is, rather than them worrying about which data to use or how to pull the data together, they simply give the task to the agents,” he said.
For example, a company might ask for an assessment of the competitive landscape for a product and the agents would pull up the top competitors, which users could then click on to see detailed attributes of each, such as taste and flavor, nutritional value, price vs value, formulation quality, shelf life and more. Their competitive stance for each attribute is then characterized, for example, as ‘high’ or ‘medium’ to help users better understand the significance of each in relation to the others.
“Previously, if a company wanted to figure out the comparative intelligence, it would buy a dashboard” and need to connect the data and then analyze it, Shrimali said.
“This agent is so well trained on this domain, you just give it your KPI and will do everything else for you,” he added. “It is so proactive you don’t need to learn a software. There is no need to learn 100 different things,” or to worry about electronic notebook systems.
The inspiration for the “all-in-one” collection came after watching companies “get sucked into completely the wrong direction” and spend years simply inputting data into electronic notebooks “only to realize the data was meaningless,” he said.
Tried and tested
Major CPG companies already are using Turing’s Luna AI and generating great results, including faster innovation and execution at market and cost savings, Shrimali said.
After collaborating with them to refine the tool, Turing is now ready to roll the product out to everyone.
Shrimali expects companies will need to join a wait list initially, but he expects they will have access within weeks.
