Every generation of food and beverage leaders reaches a moment where they had to decide what to trust. Your grandfather trusted his gut because the business was small enough and he could see everything for himself. Your father trusted the reports his systems produced because the business had grown past what one person could know.
Now it’s your turn, and deciding where to place your trust is a tougher decision than ever before given that your employees are likely already relying on a choice outside of your control. In a recent survey of food and beverage decision-makers fielded by Vanson Bourne, 44 percent of employees admitted to using AI tools without formal approval, above the cross-industry average of 40 percent.
I ask F&B executives a version of this question all the time: how is AI showing up in your operation? Most leaders think they know the scope of AI use in their business because they assume it’s only what they’ve sanctioned: a licensed tool, pilot, or a defined roll out. The real finding indicates adoption is ahead of and already going beyond what leadership approved.
Employees Are Not Out to Break the Rules
The reality is these unapproved tools are easy to find, and employees don’t wait for permission to use something that can make their job easier, and they likely won’t choose the solution that can result in better information and decision making with the potential for improved business results.
They aren’t hiding it, but they also aren’t reporting it because there isn’t an expectation they should. A reported 78 percent of food and beverage organizations cite a lack of governance as a barrier to getting real value out of AI, and I’d argue it’s a result of this dynamic of adoption outrunning oversight.
Often when I talk with smaller companies, some of the users are savvier with AI tools than the IT guy. He may have been there 25 years, is kind of old school, and doesn’t have the bandwidth to keep up with the latest tools. And then you've got some newer, younger, very sharp employees who are using AI in their private lives, and they want to use it now to make their work life easier.
I see this also as a good sign that people already believe AI can help them, but it does come with some risk.
Regardless of where unsanctioned AI usage starts within the business, the stakes are higher in food and beverage manufacturing. Much of what these companies run on supplier records, customer information, batch and lot data, and formulations, all tied directly to traceability and food safety.
The direction we're heading in is to make all traceability information available to the appropriate authorities. That's also something consumers are asking for: transparency and knowing the source of your food and what's in my food. Where you could get into governance risk is if you started trying to gather and assess traceability data to respond to inquiries using public AI tools.
For example, when someone uploads a supplier contract or a batch record into a public AI tool, they may be exposing confidential information without meaning to. They may also get a hallucinated answer back because the tool has no idea what a changeover, catchweight, or a lot code actually is.
Of those organizations surveyed, 81 percent said data quality and access are their greatest barrier to AI success, and that’s not a coincidence. You can’t govern data you can’t see, and you can’t trust an answer built on data you don’t trust.
A Closer Look Inside
I've been involved with large enterprise companies implementing supply chain planning solutions with complex regulatory and customer requirements, and when I delved into the data in their ERP system in terms of the integration, there was room for improvement which impacted the implementation timeline.
Just like ERP integrations with supply chain planning, warehouse management or transportation management solutions it’s even more critical to have complete and accurate data for AI to work properly. It all kind of comes down to that foundation. You have to have good data to work with to get good results.
None of this is an argument against AI, but rather an argument for being deliberate in your AI use and governance. Unlike general-purpose AI tools that can carry hallucination risk, industry-specific solutions are designed around what really impacts food and beverage manufacturers: areas like shelf-life constraints, lot-level traceability, batch records, inventory complexity, and production scheduling challenges.
We found that 46 percent of F&B organizations have adopted AI built specifically for their industry, and reported stronger results.
F&B organizations using industry-specific AI tools in our research saw a 29 percent improvement in forecast accuracy, compared to 19 percent for those using only general-purpose tools. On competitive positioning, industry-specific AI users reported a 39 percent improvement against 32 percent for general-purpose users.
The difference isn’t really about which AI tool is smarter. It comes down to the reality that when you feed a system clean, connected, industry-specific data, that understands your industry, it gives you something you can act on. Something you can trust.
With 83 percent of our research respondents believing their business will become obsolete without successful AI adoption, we’ve reached the moment where you must decide what to trust.
My advice? Don’t just go out and buy an AI tool and say, “I want to start using this.” Consider your business case. Ask where employees are already using AI and what problems they are trying to solve and have solved. Use that to develop an AI strategy with leadership, guidance, and governance.
With your business leaders embedded very heavily in the outcomes you're trying to achieve, you’ll quickly catch up with your employees.
Findings from the Aptean 2026 Artificial Intelligence Research, commissioned by Aptean and conducted by independent research firm Vanson Bourne. Data collected from 1,535 business leaders across five verticals in Q2 2026. Read them in full here: https://www.aptean.com/en-GB/food-and-beverage-ai-report
