A current trope in the product design community is that consumers are no longer buying a coffee maker, microwave, lightbulb or any other device. They are buying smart, connected versions of these products. As a result, software development is becoming essential to product development. Companies that have been producing physical products for generations, are now hiring thousands of software developers to ensure the intelligence added to their products is implemented correctly.

The SmartWorks suite offers a workflow to get and analyze data from devices and use that to develop and deploy AI. (Image courtesy of Altair.)

“They are quickly becoming as much a software organization as they are a manufacturing organization,” said Christian Buckner, VP of Product Management at Altair. “And just like industrial design and engineering groups, these software organizations need to correlate how product changes result in efficiency gains and ease of use. They need to understand how user behavior leads to user requirements and how to adapt to them.”

To this point, Altair developed SmartWorks Analytics and SmartWorks IoT. The former is a low-code software for people performing data science for AI applications. The latter is a low-code software used to develop IIoT applications, connect assets and build platforms.

Buckner notes that each tool has its own place within the development of the digital enterprise. However, the two software tools share a common name because in a smart enterprise IoT, data analytic and AI systems will all be linked together. “Together, these tools comprise a complete workflow for getting data from devices, analyzing that data and then developing and deploying AI models so users can harness the insight that comes from this automation and intelligence,” he said.

How SmartWorks Analytics Can Develop AI Tools

Altair designed SmartWorks Analytics to process any data workload: from small text files to data stores and data lakes. It was built to be used by technical and non-technical users to develop data analytic models without needing the background and training that would traditionally be required. As a result, anyone can have access some of the most popular open-source data analytics tools available.

“In its simplest form,” Buckner said, “SmartWorks analytics is a drag-and-drop pipeline creation and data preparation tool. It allows you to connect to dozens of different enterprise data sources, transform and join those sources, then write that transformed data to equally as many data destinations. It’s a drag-and-drop way to effectively generate the format of data that you want. It’s also high performance and cloud-based so you can do it in a collaborative way.”

Buckner notes that collecting and collating data is only half of what SmartWorks Analytics is designed to do. It is also designed to create predictive models and AI algorithms based on that transformed data. “It allows a single user to take raw data from many sources, engineer it, and then build advanced inference models without writing a line of code,” he said.

To use the model in the wild, users need to host it and provide an API to access it. This is where SmartWorks Analytics MLOps comes into play. “It provides an environment for AI models to run and be interfaced with,” Buckner explains. “It also provides a set of tools for model developers to track, monitor and continuously improve the model.”

How SmartWorks IoT Can Produce an Ecosystem of Things

SmartWorks IoT was designed to have two primary functions: to help product developers build web- and mobile-apps to complement physical products, and to help operational teams connect to industrial tools and equipment.

“At the simplest level, SmartWorks IoT acts as a cloud or on-premises centralized backend for an IoT ecosystem. It allows you to build applications which talk to assets and users regardless of where they are in the world,” Buckner said.

SmartWorks IoT also includes the tools you need to manage that backend application, such as:

Device management
Device connectivity
Data storage
Application logic engines
Serverless functions for automation
Role-based access and control
Edge orchestration

“Moving up in complexity,” added Buckner, “SmartWorks IoT also offers a rich, real-time, drag-and-drop dashboard creation tool to build IoT applications. It allows users to build different applications for different user groups. For example, for end-user customers, you can build a custom mobile app. But then the product development group could build themselves a dashboard to analyze user behaviors for requirements gathering. The operations group could build a maintenance dashboard to optimize their schedules. It can all be built and maintained using the same tool, backend and drag-and-drop interface.”

He continued by noting that one of the most complex functionalities of SmartWorks IoT is its ability to implement edge computing to connect to industrial protocols, clean up data and run AI models or automation systems.

“This tool comes with a full developer portal for monitoring and improving that automation,” said Buckner. “That gives product development groups full control to update and improve very complex code running on highly intelligent assets, like autonomous robots or cameras used for computer vision, without the risks associated with transmitting large amounts of code over the air.”

In other words, Buckner is saying that SmartWorks IoT is all about giving product development companies the ability to simultaneously improve designs, deployment, operations, manufacturing, product development and the user experience through the collection of data from users, equipment and products, and the ability to deploy software updates to whatever is connected to the ecosystem.

It Isn’t Just Smart Development, It’s Smart Companies

With the ability to implement new AI tools and develop user specific tools, engineering teams could produce numerous tools to help their entire organization better understand their products and how their individual jobs affect the outcome.

(Image courtesy of Altair.)

For instance, Eliyahu Goldratt would tell you, the goal of any company is to profitably generate money. And Buckner notes that it’s harder to do this without the marketing and sales teams being up to date on the status of the product, even during development.

“They have as much stake in R&D and how users in the market react to the product and they need to use that information to figure out how to better sell the product,” said Buckner. It goes beyond these obvious links; he notes that finance, operations and HR will need access to sales, marketing and maintenance just as much as the R&D group, so they can better serve customers and the internal organization to make the company more successful as a whole.

“What we’re seeing is the need not just to connect design, engineering and manufacturing and services with a set of tools, but instead connect the entire enterprise,” he added. “Every group in the company including sales, marketing, software development and IT are all intrinsic to the product success and should be just as connected to the flow of data information as traditional product development teams.”

This all relates to Altair’s general directive to offer numerous tools to everyone working for their customers. Currently, through Altair Units, a non-consumable token-based business model, customers have access to over 70 Altair tools. Anyone at the company can access any of these tools, from anywhere the company operates. Instead of offering each tool on a per software and per seat basis, users can access them all at a certain bandwidth.

“I think the goal for us is to build collaborative connected software products that unite these different groups, that are involved with a digital enterprise, while simultaneously making it economical as a business model as companies invest more and more into Altair’s portfolio,” he added. “The more tools we have, the more of the organization we can help, and the better it is for that organization.”