If you own or manage a manufacturing, distribution or warehousing facility it’s time to get serious about artificial intelligence. That’s because the technology is starting to mature into something that is actually generating ROI. Software and hardware vendors are rolling out AI technologies both independently and as part of their existing products and leveraging these technologies is critical for increasing productivity, reducing overhead and increasing profits.

Here are four areas to consider.

Robotics

Manufacturing and distribution businesses of all sizes are investing heavily in AI-powered robotic technologies for the shop floor. Walmart is rolling out autonomous forklifts driven by AI software. Amazon is deploying hundreds of human-shaped warehouse robots to lift and move packages as well as independent drones to monitor traffic, safety and workflow. A German company called Robco is lending out robots that perform lathe turning laser engraving and palletizing (the robots can then be reconditioned and used for other customers as needed). Other robots are being used in food processing and baking operations to independently mix ingredients and then load finished products into cartons and cases.

Some think that humanoid robots are for the movies. But that’s not the case. They’re appearing more frequently on the shop floor – alongside and in lieu of workers.

For example, 1X creates an abundant supply of labor via safe, intelligent humanoids for the manufacturing floor. Agility Robotics is pumping out tens of thousands of humanoid warehousing robots. Boston Dynamics now makes an electric powered version of its humanoid robot Atlas. Apptronik’s Apollo is a general purpose humanoid robot designed to work in the plant and the office.

Unbox Robotics offers a bunch of AI “swarm robotics” to “accelerate the parcel sortation and order fulfillment to facilitate efficient express logistics operations delivering seamless end customer experience.” United Robots produces AI-powered robotic cleaners, disinfectors and carriers that do everything from removing microbiological contaminants to safeguarding workers to tugging heavy loads or carts. Other humanoid robots designed to perform all sorts of industrial tasks – and winning awards and – are now available (or coming soon) from companies like Figure AI and Sanctuary AI.

The robots on the manufacturing floor are controlled through the cloud using platforms provided by Nvidia (the chipmaker refers to the new platform as “a general-purpose foundation model for humanoid robots.”) and the aforementioned Agility Robotics, which the company says can “command a robot army, say, to start moving bins to a conveyor belt at a particular time.” Not sure I like the sound of that, but OK.

Internet of Things

Industry Week’s VG Govindarajan and Venkat Venkatraman write that smart companies like Rolls Royce are using “fusion technology” being powered by AI is turning yesterday’s equipment into thinking, intelligent machines.

“Tractors have become “farm-based computers” linked to precision agriculture cloud systems,” they write. “Rolls-Royce offers “Power by the Hour” agreements for its aircraft engines. This approach ensures that Rolls-Royce is responsible for the maintenance and performance of its engines, with costs linked to the hours that the engines are in operation.”

Firms with complex processing applications – particularly larger power generation, oil and gas, chemical and food and beverage manufacturing companies – are now implementing AI platforms like UptimeAI which uses sensors to monitor, analyze, evaluate and alert when there are disruptions, malfunctions or other problems that can create anomalies, loss of efficiency and safety issues.

Other software firm and hardware manufacturers are partnering to create Internet of Things (IoT) sensors that are using AI technology and models. The sensors are then installed on manufacturing equipment for the purposes of “making hardware smart” using audio, video, vision, temperature and pressure sensing.

3-D Manufacturing

A variety of companies – like Fathom Digital Manufacturing , Protolabs and Stratasys – are using AI to create 3D printing solutions to “dramatically accelerate the product development cycle” that allow manufacturers to design, create, test and evaluate prototypes in “a matter of days, rather than weeks or months.”

Consolidating Data Into Large Language Models

For AI to work, large databases must be built from disparate information.

One company called EthonAI has launched a new software platform that they say “is designed to provide managers and engineers with a “sixth sense” to follow what’s happening within their factories more closely.” The platform brings together data from multiple sources and then uses AI to make recommendations for management.

“With AI tools, it can be easier for companies to identify patterns that produce poor quality products and implement changes quickly,” writes Rae Hartley Beck. “When a greater percentage of final products pass quality control, less waste is produced and more value is created for companies.”

Another company bringing data under one roof – Sight Machine – recently announced a partnership with Siemens AG to create a manufacturing data platform for collecting, contextualizing, and analyzing all types of manufacturing data to improve production. Sight Machine’s software “analyzes the entire history of production runs to determine the best settings for the current conditions.”

Microsoft announced a number of initiatives to bring data under one roof using “Copilot templates” so that their customers can “create their own copilots.” The idea is for employees to use natural language queries to retrieve data – like asking the large language model why a machine is breaking and getting insights to help resolve the issues faster.

Manufacturers in the auto industry are reportedly investing in internal AI applications to help improve production scheduling, minimize line stoppage, testing chemical mixtures, aligning data from different sources to better communicate with suppliers and customers and sourcing materials based on real time availability around the world to reduce supply chain bottlenecks.

In the back office, cloud computing platform ServiceNow has released financial, customer service and operations software designed specifically for manufacturers that uses AI to do things like automatically recommend replacement parts for customer maintenance, automatically manage customer exceptions, communicate independently with suppliers and autonomously log a problem and deploy a field technician with the necessary knowledge, skills, and equipment to manage the task. Other manufacturing software makers like Epicor, SAP and Sage are rolling out similar AI based functions.

OpenAI’s ChatGPT was released in November, 2022. It created a flurry of activity, jumpstarted a tidal wave of investments, launched countless startups and forced every software and hardware company to double down on their AI offerings. Less than two years later, manufacturers are now starting to see the results of these efforts. And we’re just getting started.

FOXBORO — If the Patriots’ quarterback battle heats up in the final month of training camp, mark Thursday’s 14-13 loss to the Eagles as the starting point of the true competition.

Patriots rookie quarterback Drake Maye impressed and led his team on two scoring drives in his first considerable playing time of training camp.

Here’s what we learned as the Patriots fell to 1-1 in the preseason.

1. QB competition got interesting

Patriots veteran quarterback Jacoby Brissett didn’t do anything to help himself, playing three drives to start Thursday’s game. The Patriots went three-and-out on his first series. Then he led the offense down the field on an eight-play, 31-yard drive before throwing an interception on a target to tight end Austin Hooper, who was too well-covered by Eagles cornerback Avonte Maddox for the pick.

Brissett finished just 3-of-7 for 17 yards with the interception.

Then Maye brought some excitement. His first drive ended with a 51-yard field goal from kicker Joey Slye. Maye’s best pass was a 12-yard completion to rookie wide receiver Javon Baker on third down after a 6-yard scramble to make the first-down conversion manageable. The drive stalled when Maye threw a deep ball down the left sideline out of bounds to Baker.

Maye scored on a 4-yard run to cap off his second drive. He kept the ball on a zone-read to scamper to his right for a score. Maye hit rookie wide receiver Ja’Lynn Polk for a 5-yard gain. Polk should have been stopped for no gain, but he made two Eagles defenders miss to pick up extra yardage. Running back Kevin Harris ripped off a 14-yard run with solid blocking from the entire offensive line. Maye also hit running back JaMycal Hasty on a 23-yard catch and run.

Maye went three-and-out on his third drive, which came in the third quarter. He delivered a perfect deep ball to Baker, but the rookie wideout couldn’t hold on as he fell to the ground on the diving attempt.

Maye took a sack to cap off his fourth and final drive. Harris chipped Eagles pass rusher Nolan Smith, but left tackle Vederian Lowe still couldn’t get in his way before the sack. Maye showed good decision making, throwing away two passes on the drive. The Patriots picked up 15 yards on a defensive pass interference penalty on the series, and Maye’s lone completion went for no gain to Polk.

Maye finished 6-of-11 for 47 yards with four rushes for 15 yards with a touchdown.

So, what does this mean for the QB competition? We’ll see. But Maye hasn’t received a single first-team rep in training camp. Now that he’s proven himself in a game, perhaps that changes and the starting battle actually begins to take shape with a few weeks left until the regular season.

2. OL remains unchanged

The Patriots started Lowe at left tackle, Sidy Sow at left guard, David Andrews at center, Mike Onwenu at right guard and Chukwuma Okorafor at right tackle, which means there were no adjustments after a rough showing during a joint practice with the Eagles on Tuesday.

Maye did not receive an entire backup offensive line when he entered the game, however. First, Nick Leverett replaced Andrews at center. Then on his second drive, rookie Layden Robinson came in at right guard, and rookie Caedan Wallace came in at right tackle. For Maye’s fourth drive, Michael Jordan replaced Sow at left guard.

Overall, the offensive line performed OK outside of Lowe, who let up the sack, a pressure on an incompletion from Brissett and was flagged for a false start. It must be taken into account that they were not blocking the Eagles’ top defensive line.

3. Pressure without Judon

Let’s use that same caveat: The Patriots’ defense was not playing against the Eagles’ top offensive line, but they were causing fits for Eagles backup quarterback Kenny Pickett.

After trading Matthew Judon to the Falcons for a 2025 third-round pick on Wednesday night, defensive ends Keion White and Deatrich Wise, cornerback Isaiah Bolden and outside linebacker Josh Uche all had first-half sacks.

White, Wise and Uche will all be key in replacing Judon on the edge this season.

Defensive tackles Jeremiah Pharms and Trysten Hill, outside linebacker Oshane Ximines and safety Joshuah Bledsoe also brought pressure in the first half.

Ximines sacked Eagles QB Will Grier in the fourth quarter, as well.

4. Kicker competition

Incumbent kicker Chad Ryland still hasn’t gotten a shot at a field goal in the preseason, while Slye, his competition, is 3-of-3. Slye handled one kickoff and made field goals of 51 and 46 yards.

Ryland was 1-of-1 on an extra-point attempt. He handled two kickoffs.

The competition has been tight throughout the summer.

Up

QB Drake Maye

Was he perfect? No. Did he give the Patriots a spark? Absolutely. Maye only received six offensive snaps in Week 1 of the preseason. He showed flashes of why he was the third overall pick on Thursday night with his arm and his legs.

LB Raekwon McMillan

McMillan, coming off of his second season-ending injury in three years, looked like a classic Patriots linebacker. He was a force in the run game with eight tackles with a tackle for loss. He also assisted on another tackle for loss and contributed a special teams tackle.

DE Keion White

White is the top player who needs to step up in Judon’s absence. He sacked Pickett and brought pressure on another 3-yard desperation scramble from the backup QB.

Down

LT Vederian Lowe

Lowe is the current top left tackle. We’ll see how much longer that lasts after Thursday night, when he let up a sack, a QB hit and was flagged for a false start. Third-string offensive lineman Atonio Mafi also struggled, allowing two sacks.

CB Marcellas Dial

Dial was a favorite target of Eagles quarterbacks. The rookie, who’s battling for a roster spot, appeared to let up seven catches for 87 yards. He also allowed a two-point conversion.

WR Javon Baker

The level of difficulty was high on the pass from Maye, but Baker should have hauled in the rookie QB’s deep shot. Baker also ran a route almost completely out of bounds.

Originally Published: August 15, 2024 at 9:58 p.m.