Manufacturing’s AI ambitions confront the reality of data and organisational change
The manufacturing industry has no shortage of ambition when it comes to artificial intelligence. Across factories and industrial organisations, companies are exploring how advanced analytics, automation and increasingly sophisticated AI systems could improve productivity, strengthen resilience and transform operational decision-making. Yet the challenge of turning that ambition into measurable business value remains considerably more complicated than deploying the technology itself.
That tension provided a central theme for the Manufacturing Data Summit (MDS), held in London on October 6, 2026. Bringing together manufacturing leaders, data specialists and technology experts, the event examined the practical realities of industrial digital transformation, from establishing reliable data foundations to addressing cybersecurity, organisational accountability and the difficulties of scaling AI beyond isolated applications.
Organised by Manufacturing Today, the summit brought together perspectives from organisations including Coca-Cola Europacific Partners, DS Smith, British American Tobacco, Medik8, Johnson Matthey, Mars and Babcock International Group. The programme combined keynote presentations, panel discussions, practical case studies and smaller roundtable conversations, to encourage open exchanges about the challenges manufacturers face.
What emerged was a picture of an industry increasingly aware that technological capability is advancing faster than many organisations’ ability to deploy it effectively. AI may offer new ways to analyse industrial information and support operational decisions, but its value depends on the quality of the underlying data, the clarity of organisational responsibilities and the willingness of employees to adopt different ways of working.
Moving beyond the AI pilot
One of the summit’s central questions concerned the gap between successful technology demonstrations and enterprise-wide transformation. Manufacturers have invested considerable effort in exploring AI applications, but translating those experiments into repeatable improvements across multiple facilities requires more than technical proof that a system works.
The conference programme challenged participants to consider where digital investment should be directed, what separates a successful pilot from an initiative that never progresses beyond experimentation, and how organisations can establish a credible business case for scaling new capabilities.
An important distinction running through the discussions was between demonstrating what technology can achieve and establishing the organisational conditions required to sustain that achievement. Industrial AI depends on access to information from equipment, production systems and wider business applications. However, simply collecting that information does not make it suitable for analysis. Manufacturers must also establish consistent definitions, understand the context in which data is generated and determine who is responsible for maintaining its accuracy.
These requirements become more demanding when organisations operate multiple facilities with different systems, production processes and management structures. A solution developed for one factory may not transfer easily to another unless the underlying architecture, governance arrangements and operational requirements have been considered from the outset.
The importance of organisational alignment became particularly apparent during discussions about the relationship between information technology and operational technology. While IT departments typically prioritise network security, architecture and enterprise standards, operational teams must ensure that production equipment remains safe, reliable and available.
Neither perspective can be treated as secondary. Successful convergence requires manufacturers to establish shared objectives while recognising that responsibility for technical infrastructure and accountability for operational outcomes are not necessarily the same thing.
A recurring theme was the importance of assigning clear accountability to the business functions that ultimately depend on industrial information. Shared responsibility across IT, engineering and operations can encourage collaboration, but it can also create uncertainty when nobody has explicit ownership of the results.
The discussions also highlighted the difficulties created by organisational silos. Skills and specialist expertise may already exist within a manufacturing business, but they are not always available to the teams or facilities that need them. Overcoming those boundaries can be as important as acquiring new technological capabilities.
Turning industrial data into operational value
The importance of data governance was another recurring theme, particularly as manufacturers begin introducing AI systems capable of interpreting information and recommending actions across increasingly complex operations.
Traditionally, governance has often been associated with compliance, documentation and policies governing how information is stored or accessed. The discussions at MDS presented a more operational perspective, examining how unreliable information can create direct costs through slower decisions, duplicated effort, unnecessary inventory and inconsistent planning.
When employees cannot trust the information available to them, they frequently resort to checking figures manually or maintaining separate spreadsheets. These workarounds may solve immediate problems, but they can also create competing versions of operational reality, making it harder to coordinate decisions across departments.
The arrival of agentic AI introduces additional questions. Systems that previously provided information for employees to interpret are increasingly being designed to perform tasks or initiate actions. This changes the significance of access permissions, data ownership and the controls governing what software is authorised to do.
The distinction between allowing an AI system to examine information and permitting it to act upon that information is particularly important in manufacturing, where incorrect decisions could affect production equipment, material availability or operational continuity.
Participants also considered whether manufacturers should collect and centralise every available piece of factory data. Although greater connectivity can improve visibility, moving large volumes of high-frequency equipment information into centralised systems can increase costs and introduce additional security considerations without necessarily improving decisions.
A more selective approach, based on clearly defined operational requirements, offers an alternative. Data can remain close to production equipment where immediate processing is necessary, while information required for broader analysis can be made available to enterprise systems.
Many manufacturing assets were installed before modern connectivity and cybersecurity requirements became established, making legacy equipment particularly difficult to integrate with contemporary data platforms without additional interfaces, gateways and protective measures.
These challenges reinforce the importance of designing digital infrastructure around specific business outcomes rather than pursuing connectivity as an objective in itself.
The summit also explored what becomes possible once organisations establish more reliable and accessible information.
One operational AI case study examined how a growing consumer products business had moved beyond spreadsheet-based inventory planning by combining information from enterprise applications within a common data environment.
The approach supported more detailed forecasting of stock requirements, replenishment schedules and future inventory positions. AI-assisted development also enabled the creation of tailored planning tools and systems capable of identifying changes in operational information and bringing them to the attention of relevant employees.
However, the experience demonstrated that technological capability does not automatically create user confidence. Employees need to understand how systems reach their conclusions, particularly when AI-generated recommendations replace familiar spreadsheets and established manual processes.
Trust must therefore be developed alongside technical capability. A system that produces useful recommendations but cannot explain the information and calculations behind them may struggle to gain acceptance, even when its underlying models are reliable.
Building the foundations for lasting transformation
Beyond the immediate challenges of AI deployment, the summit programme considered how manufacturers can ensure that digital transformation strengthens rather than complicates established operational disciplines.
As factories become more connected, cybersecurity is taking on greater importance, with manufacturers having to balance the benefits of greater data availability against the risks of exposing production environments to additional systems and potential vulnerabilities. The programme examined emerging regulatory pressures, supply chain exposure and the responsibilities of senior leadership in establishing effective cyber resilience.
Digital twins provided another opportunity to examine the difference between technological promise and practical application. Rather than treating simulation and virtual modelling as universal solutions, the session focused on identifying circumstances in which these technologies can deliver measurable improvements to maintenance, production flow and operational performance.
Similar questions informed discussions about lean manufacturing. Digital tools can improve visibility and support process optimisation, but they can also introduce complexity if their implementation is disconnected from established improvement objectives.
The underlying challenge is to ensure that technology contributes to operational performance rather than becoming another layer of activity that employees must manage.
The event concluded with a series of interactive roundtables examining the practical challenges of scaling AI, building a culture capable of supporting digital investment, prioritising valuable information and translating data strategies into measurable manufacturing outcomes.
These conversations formed part of a wider emphasis on collaboration and knowledge sharing, recognising that many organisations face similar difficulties even when their technological maturity, manufacturing processes and commercial priorities differ.
Reflecting on the event, Dr. Yaniv Naor Nahoum, Founder of N-BITE Advisory, highlighted the value of bringing different perspectives together. “The Manufacturing Data Summit brings together industry leaders, innovators, and decision-makers who are shaping the future of data, AI, and digital transformation in manufacturing. What makes the event especially valuable is the quality of the conversations: practical, relevant, and focused on the real challenges organizations face when turning strategy into execution.”
That emphasis on practical discussion was also reflected in the wider conference programme, which repeatedly returned to the need for clearer ownership, stronger governance and closer cooperation between technical specialists and operational decisionmakers.
The implications extend beyond AI. As manufacturers adopt increasingly connected and intelligent systems, the distinction between digital strategy and operational strategy becomes harder to sustain. Decisions about information architecture, cybersecurity, workforce capabilities and technology investment increasingly influence the ability of factories to improve productivity and respond to changing business requirements.
For manufacturers seeking to make progress, the strongest message from the recorded discussions was the importance of beginning with a clearly defined operational objective, establishing accountability and measuring the results before attempting to replicate an approach more widely.
The Manufacturing Data Summit demonstrated that the industry’s next phase of digital transformation will depend as much on organisational capability and disciplined execution as on advances in artificial intelligence. The technologies may be changing rapidly, but sustainable improvement still requires manufacturers to understand their processes, trust their information and connect investment to outcomes that matter.
