• Day one at TechEx North America: operational changes for transformation projects in 2026.
• Multiple show tracks each bring detailed focus.
• The issues facing decision-makers in the enterprise today.

Day one at TechEx North America put the case for constraints on enterprise AI – not to limit LLMs’ abilities, but in the form of questions around whether the enterprise has the operating model, data layer, and governance to make it work.

The AI Strategy track opened with a keynote panel on migrating ‘vanilla’ AI assistants to autonomous agency, covering system architecture, governance, and the points at which humans still need to be in decision chains.

If anything, the opening sessions defined the many elements for discussion that played out over the course of the day. The message that enterprise AI is a permanent part of the architecture means considerations are in play that sometimes have little to do with technology: risk, finance, and change management are what are at the front of decision-makers’ minds.

Lutz Beck, chief information officer for Europe and APAC at Daimler Truck Financial Services, described the evolution from smart chatbot to independent agent. His session was billed around humans and machines building what the show schedule termed a ‘real-time enterprise’. In such an organisation, AI is treated as part of everyday work, and sessions later in the day took the same line – there was a genearl acceptance of agentic AI’s role in many business functions.

Speakers from IBM, eBay, Deloitte, SAP, BMO, and HPE dealt with many aspects of the practical design of enterprise AI deployments, with the details being filled in for each of the contextual layers: the pros and cons of autonomous decision-making, the planning required at C-suite level, and the range of financial outcomes that the business might expect to see.

There were lessons presented to the show audiences around both technical architecture of digital transformation, and the operating expenses that may be incurred from decisions taken. Token costs was a theme across several sessions, presented as a problem for management, rather than one that can be solved by engineering. As the big AI providers spread token-based billing from the enterprise price tiers down towards smaller-scale AI users, the overall sentiment was that subscription models – “all you can eat” for a fixed monthly fee – are gradually walking off into the West Coast sunset.

The cost of any business process is important, and when AI-powered decision making is being debated, it’s undeniably important. Once an AI agent is expected to take action rather than, for example, draft recommendations, then there are audit trails, lines of responsibility, and budget allocation issues to be sorted through. Getting the inside knowledge and lessons born of experience from some of the biggest businesses in the world is what this event delivered.

The Data at Scale sessions supplied some of the more complex parts of the discussion around digital transformation in 2026. Human oversight, AI coworkers, data products, data marketplaces, and data science were discussed as conditions for (and must-have elements of) successful AI deployments.

The agenda placed human oversight considerations next to sessions on how best to monitor the new generation of our AI coworkers. Can organisations be sure that their data is sound, before basing their strategy on it? And in the case of agentic AI, can LLM-based systems remain financially viable after they enter live work from isolated pilot projects?

The Data, Platforms and Systems track dealt with storage pipelines, master data, analysis engineering, governance, orchestration, and employee upskilling. Here, the sessions were more technical, but always in the context of commercial lessons; learned and shared from the stage.

It was a pleasant surprise to discover that the Digital Transformation track of the show gave serious airtime to matters not technological but human. The day-one track touched on stalled transformations and those workflow automation projects that are mid-transition to agentic intelligence. Its sessions considered why transformation programmes might fail despite the availability of cloud platforms, the huge choice of automation tools, and the many AI options.

The more convincing answers presented were often about getting the simple things right: teams need clear workflows, clear accountabilities, and systems of change that can become the pillars of a responsive business.

There were sessions on employee engagement, customer experience frictions (and their effects on the bottom line), and on how to achieve financial return on AI and automation projects. The overriding sentiment was that the age of novelty, ushered in by the emergence of gen AI, is well and truly over.

In 2026, it’s all about implementations that make financial and operational sense, and making sure the workforce is prepared and waiting to go.

It’s sometimes difficult to choose highlights from a large event, especially given that so much takes place in different areas of the conference at the same time. But one session’s contents especially was relevant to just about every attendee: The AT&T case study on Microsoft Copilot centred around real-world roll-out, user behaviours, and adoption of the ‘modern-day Clippy’ by the enterprise. Sure, Copilot can summarise a meeting or draft a note, but can a large workforce actually change its working methods with help from the smart O365 tool? How do decision-makers differentiate employee activity from value? Having a starting point for discussion in the form of the ubiquitous O365 meant that this was one of the more popular destinations on the show floor today.

The Intelligent Automation programme extended the theme into automation economics. The day-one sessions covered ROI in automation projects, agentic automation, the limits of traditional robotic process automation (RPA), robotics in enterprise settings, AI-powered OKRs, ‘IT operations at AI speed’, Python governance, and real-world deployments of new code. The phrasing varied from session to session, but in general, automation’s quick-wins come when it removes friction or shortens a process reliably and at scale. Automation is weak when it is simply placed on top of badly designed workflows.

Classical robotic process automation often worked by imitating a user inside existing systems. Agentic automation promises a more flexible range of actions, but demands authority, evidence, and human supervision. A bot that copies data between fields can be governed deterministically, yet an agent that decides which process step should happen next needs clearer policy boundaries.

The strongest theme in the AI, data, transformation and automation tracks at TechEx North America was convergence. Although it’s tough to summarise an entire day’s discussions, the AI track talked about agents, the data track about governed foundations, and the automation track talked about ROI and operational control.

The day’s sessions pointed to enterprise AI moving into the machinery of management, but at the present moment in the technology’s evolution, it depends on several existing elements that need to be given new focus: platform discipline, a coherent data layer, realistic economics, and the human element of accountability and sanity-checking.

You can explore many of the issues in today’s spotlight on the TechForge portfolio of sites, which features interviews and articles with many of today’s speakers and exhibitors at TechEx North America, 2026.

(Image source: TechForge)

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.