Nextmv raises $8 million to optimize decision-making within applications

Nextmv, a provider of optimization and modeling tools that enable developers to automate decision-making within an application, today revealed it has raised $8 million in a series A round of funding.

The tools Nextmv provides trace their lineage back to route optimization research the company founders conducted while working for Grubhub, Nextmv cofounder and CEO Carolyn Mooney said.

Based on that experience, they realized developers would benefit from a set of tools to make the algorithms required to, for example, optimize logistics for a fleet of trucks, Mooney said. Those offerings are Hop, an optimization and modeling tool, and Dash, a tool for experimenting with simulations. The company also has a free Fleet Routing Model on Nextmv Cloud available in alpha.

FirstMark Capital led this round, with participation from original seed fund investors Dynamo Ventures, 2048 Ventures, Atypical Ventures, and Greenhawk Capital. Individual investors joining the round include GitHub CTO Jason Warner, Stripe COO Claire Johnson, Twilio VP Rick Robinett, and Seamless founder Jason Finge.

Historically, achieving route optimization has required an individual with a lot of specialized knowledge. Lately, many organizations have been investing in various data science initiatives to achieve that goal using software. The tools Nextmv provides are intended to allow developers to model decisions in a way that makes it easier to optimize them, Mooney said.

Beyond fleet optimization, other use cases for decision-making tools include cloud resource allocation, inventory management, and portfolio management, Mooney added. They could be employed to optimally connect sales leads to appropriate agents, pickers and packers through a warehouse, or applications across a fiber network. Because not all issues should be addressed in the order they arise, organizations need a way to optimize decisions based on business goals such as profitability or the satisfaction of their most important customers. Each of these use cases shares the same challenges, namely routing to find the most efficient path, packing to determine what load can be carried, and clustering to make sure like things are grouped together.

Naturally, a wide range of business processes are inefficient simply because they are generally too complex for even a specialist to consistently optimize. “There’s a lot of room for improvement,” Mooney said.

During the economic downturn brought on by the pandemic, the need to optimize business processes has become more acute, either because companies had to reduce the size of their workforce or, in some cases, are growing at a rate they can’t sustain with hiring in-demand specialists. In theory, any number of applications should be able to optimize processes, but the underlying decision-making models for each unique process have been difficult for developers to construct, Mooney said.

Alternatively, organizations might opt to rely on a packaged logistics application, for example. However, customizing a packaged application to meet a business’ specific needs can be both challenging and expensive. Developers need access to a set of decision-making tools, libraries, and models they can easily apply across multiple use cases, Mooney noted.

Data scientists are studying a wide range of processes in the hopes of applying AI models to optimize them. However, not every process needs a complex AI model to optimize it. In many cases, the right tool in the right hands of an application developer might solve the same problem at a much lower cost.

By VentureBeat Source Link