Operators Demand Real-Time Data to Increase Uptime

Six in 10 manufacturers (59%) report actively using smart manufacturing technologies to support operations, according to Rockwell Automation’s recent report. The 2026 State of Smart Manufacturing  survey also cites that only 18% of respondents remain in pilot mode. While this survey is not centered on the food segment, recent surveys point to more real-time tools at the operator level, machine learning, and more automation for plants.

The transition to digital manufacturing in the food segment is a long one. The challenges in operations are many, including numerous changeovers per shift, multipack formats, increased SKUs, and limited operator skills. However, digital twin technology and data collection at the operator level may prove to be a step-change in productivity – a golden age of food production – while innovative machine design continues to increase uptime and throughput.

The golden age of food production

Food giants and mid-size producers want flexible manufacturing. Operators face significant challenges on the plant floor, as increases in stock-keeping units (SKUs) lead to complex changeovers and scheduling.

“In plants running multiple lines, the priority is to keep each stage of production aligned, so capacity is not lost between processing, seasoning, weighing, packaging, and downstream handling,” says Steve Johnson, Divisional Sales Manager, North America at TNA Solutions.Steve Johnson, Div. Sales Mgr., North America, TNA Solutions, says quick changeover depends on removing unnecessary decisions from the factory floor. Image courtesy of TNA Solutions

Data-first approaches and standardized automation are the new norm in today’s plant. 

“Plants are improving throughput and flexibility by optimizing whole-line performance and designing lines that are both modular and data-rich,” says Dave Reynolds, Business Development Manager of Snacks, Cereal and Protein Solutions at Bühler Group.

“Line balancing plays a central role,” adds Johnson. “Controlled distribution into the weighing and packaging section helps prevent overfeeding, starvation points and product build-up, all of which can create micro-stoppages.”

Micro-stoppages on packaging and processing lines are becoming more transparent. “Operational data is often trapped and stored for limited periods in historians, locked behind legacy protocols, or isolated in systems that were never connected to any network,” adds David Ariens, Founder at the IT/OT Insider. “Data that's perfectly adequate for day-to-day operations — an operator sees a sensor flatline and knows to ignore it — is not adequate for analytics.” 

Digital twin technology and data modeling are making inroads to maximize uptime by feeding raw data into models. Targets on the plant floor include packaging and processing line configurations, bottlenecks, and Overall Equipment Effectiveness (OEE) challenges.

SmartSights provides digital twin technology for packaging and other manufacturing applications. “Regression models shine since they can handle different variances and manage outliers,” said Marc Bertrand, Dir. of Industry Solutions, SmartSights, in a Control Systems Integrators Assn. webinar in late 2025.

“Many times operators believe a filling machine on the packaging line is the bottleneck, and we’ve run regression models and found the filler is not the true bottleneck on the line,” said Bertrand.

The same applies to processing lines. “Digital twins are becoming increasingly practical for simulating capacity changes, recipe shifts, and equipment interactions before implementation,” adds Reynolds from Bühler Group.

Recent advances in modeling using unstructured data in processing are pointing toward real-time data analytics for operators. Companies recognize the potential for targeting anomalies, micro-stoppages, and adjusting line configurations during baking, cooking, or mixing processing.

A step-change advance could be coming soon. Companies like HighByte, Litmus, and TwinThread (now part of AVEVA) are winning over managers and executives already sold on eventual AI-based strategies to take unstructured data, pipeline plant-floor data into their platforms, and add context.

The modeling platform eventually standardizes the data, exposes it to AI and machine learning, and can start iterating with an experienced operator or production line manager. “Companies can start to drive interesting insights that help operators on the factory floor so that they can start to have conversations with the data,” says Aron Semle, CTO at HighByte, Inc.

Machine suppliers are adding tool-less features to increase repeatability for operators experiencing four to five changeovers per shift. Pictured above is the TNA robag® Quantum.Image courtesy of TNA SolutionsAn experienced operator is needed to iterate on the right AI prompts. Semle provides a prompt example of a “human-in-the-loop and working with the modeled data:

1. This machine doesn't sound right.
2. Can you look at the historical data to see what's going on?
3. Give me the top three suggestions on what it could be and what to look at?

IN ACTION - AI prompts at the operator level

Older workers are retiring, and companies want to retain tribal knowledge. However, transferring tribal knowledge has been a significant challenge amid evolving technology and operations, so companies are evaluating frontline tools with AI components. 

Crest Foods, a dry food packaging company in Illinois, did just this by adopting Line Lead Champion from QAD Redzone, which is part of the company’s ChampionAI technology.

The AI-powered agent allows operators to view production lines, safety actions and quality on a tablet. “It's the shift summary that has been the thing we've latched onto at Crest Foods,” says Jared Stumpenhorst, Operations Mgr. at Crest Foods.

The company has multiple lines with horizontal form/fill/seal machines, and web failures with the film material are a constant challenge. The problem with film forming the stand-up pack is an issue that occurs “four times more frequently than any other issue in the multiple plants,” according to Stumpenhorst.

“So we went into the AI agent and said, 'Can you help us solve web failures?'” says Stumpenhorst. “The AI prompt provided us with the top four causes of web failure and directed us to three to five actions for each of the causes.”

The AI tool delivers answers quickly and provides context. “What was interesting about (the AI responses) is the titles of those recommended actions did not even explicitly state web failure is a term; these answers were all related to web failures, and it was 100% spot on," says Stumpenhorst.

Optimizing changeovers

A leading consumer trend in snacking and nutrition is experimentation. Food companies have responded by introducing new flavors, especially for millennials and Gen Z.

Suzy Badaracco, President at Culinary Tides, Inc., spoke at the International Dairy Deli Bakery Association show in early June and discussed "Hybrid Meals" and other flavor trends. Culinary Tides is a consumer forecasting company, and Badaracco cited that “experimentation remained strong post-COVID and citrus, chilies, heat, and spice are popular.”

These consumer trends can be indirectly observed on the plant floor, as changeovers increase during shifts due to more SKUs and shorter production runs. In response, machine suppliers are innovating by introducing new features to enable quicker changeovers.

Scheduling is moving closer to real-time, since teams are using plant data, inventory, demand signals and line performance information. Pictured above is the Nutrex 7 Series extruder.Image courtesy of Bühler Group“OEMs are enabling faster changeovers through modular, hygienic, and operator-friendly designs with fewer manual adjustments, guided set-up, and automation-ready features,” says Reynolds from Bühler Group, a supplier of processing equipment. Reynolds notes they have added tool-less features such as quick-release clamps, hand-adjustable guides, slide-in/slide-out components, captive fasteners, color-coding and hygienic access.

Quick changeovers are of prime importance, and so is accuracy. “Tool-less features also reduce the risk of incorrect reassembly, make cleaning more consistent and help less experienced operators carry out routine tasks with greater confidence,” adds Johnson at TNA Solutions. 

Operators are also benefiting from new approaches to validating changes during changeovers, such as Turck’s BL Ident RFID system. The tool won’t allow a packaging machine to start if the new product configuration is not correct. Turck embeds an RFID chip into miniature data carriers that are flush-mountable in metal.

“The latest tracking and verification technologies help reduce reliance on manual checks and give operators clearer confirmation that the line is set up correctly,” says Johnson.

Cheaper costs on sensors and automation are advancing these trends. “The broader trend is the combination of mechanical simplicity with digital validation and automation,” says Reynolds. “Guided recipes, sensors, digital checklists, and operator guidance can confirm that the correct parts are installed properly and that each step is completed in the right sequence.”