Call Us
.png)
We reviewed ten AI reports and publications from leading AI organisations and global consulting firms. The lessons? Systems thinking, workflow design, data governance and intentional adoption will determine whether organisations get value from AI in 2026 and beyond.
AI agents are being presented as the next major leap in business technology.
They can analyse information, coordinate activities, access systems and complete increasingly complex workflows.
But after reviewing major reports from OpenAI, Anthropic, McKinsey, Accenture, BCG, EY, Bain and IBM, we saw that the most important lessons tend to fall into four main categories:
AI use is becoming widespread.
OpenAI reports that organisations are moving from occasional experimentation towards repeatable AI-enabled workflows. McKinsey also found that most organisations now use AI in at least one business function.
The real divide is no longer between organisations that use AI and those that do not.
It is between organisations using AI as an occasional productivity tool and those embedding it into the way important work gets done.
AI projects often begin with a product demonstration.
Someone sees an impressive tool, launches a pilot and then starts searching for a business problem it might solve.
The stronger approach is to begin with a measurable operational outcome, such as:
Anthropic and BCG both emphasise the importance of concentrating on clearly defined, high-value use cases rather than spreading investment across dozens of disconnected experiments.
Licence activation, employee training and hours theoretically saved may show participation, but they do not prove business value.
A meaningful AI business case should measure whether the project improves cost, quality, capacity, risk, cycle time or customer experience.
This is where Via Technology begins its digital transformation work. We help organisations clarify the business problem, define the future state and determine whether the right solution is AI, conventional automation, process redesign or a combination of all three.
The objective is not to deploy more technology.
It is to produce a better business outcome.
Sources: OpenAI, The State of Enterprise AI, Anthropic, The 2026 State of AI Agents, McKinsey, The State of AI, BCG, Agents Accelerate the Next Wave of AI Value Creation
This is perhaps the strongest shared conclusion across the reports.
Accenture, Bain, IBM and McKinsey all argue that organisations need to redesign work rather than attach AI to existing processes.
Most inefficient workflows are not caused by one missing software feature. They are caused by combinations of:
AI can make these processes faster. Unfortunately, it can also make the confusion faster.
Before implementing AI, organisations need to understand what triggers the process, what outcome it should produce, where decisions occur, who owns the result and which exceptions require human judgement.
Generative AI is commonly used to draft an email, summarise a meeting or prepare a document. Agentic AI offers a larger opportunity.
An agent may identify a trigger, retrieve information, assess it against predefined criteria, prepare a recommendation, update another system and escalate an exception.
The question therefore changes from: “Can AI make this task faster?” to: “How should this entire outcome be delivered when people, software and AI can each perform different parts of the work?“
McKinsey and IBM both describe a future in which people and agents work together across more fluid organisational structures. That will change how work is allocated, where decisions are made, how performance is measured and which roles need to be redesigned.
Via Technology helps organisations map how work really happens, including the spreadsheets, workarounds and informal decisions that rarely appear in official process documentation. We then help simplify workflows, clarify ownership and design the right combination of people, process, automation and AI.
Because successful AI embedding does not begin with the agent, that’s only the last step
The real work begins with designing the System around it.
Sources: Accenture, The New Rules of Platform Strategy, Bain, Building the Foundation for Agentic AI, IBM, Agentic AI’s Strategic Ascent, McKinsey, The Agentic Organization
AI agents need accurate information, appropriate permissions and reliable access to business systems. For many organisations, this exposes an unexpected and immediate problem:
Their information is fragmented across platforms, private spreadsheets, email inboxes and undocumented workarounds. An AI agent cannot reliably coordinate a process when the organisation itself cannot identify the correct source of information. Bain and Accenture both highlight the importance of modern architecture, connected platforms, accessible data and strong identity management.
AI autonomy is not an all-or-nothing decision. An agent may be permitted to:
The appropriate level depends on the consequences of error, the reliability of the data and whether the action can be reversed.
As AI takes on more responsibility, ownership becomes more and more important. Every AI-enabled workflow still needs:
Organisations must assess their technology environment before adding another tool. This includes reviewing platforms, integrations, information flows, permissions, reporting needs and implementation risks.
We help organisations strengthen the digital foundations required for AI to operate reliably, rather than adding another disconnected product to an already fragmented technology stack.
Sources: Bain, Building the Foundation for Agentic AI, Accenture, The New Rules of Platform Strategy, IBM, Agentic AI’s Strategic Ascent
BCG applies its established 10/20/70 principle to AI transformation:
The exact percentages will vary, but the message is clear: AI changes how people complete work, make decisions, share knowledge, evaluate quality and understand their responsibilities.
A technically impressive solution can still fail when employees do not trust it, understand it or know when to challenge it. Successful implementation requires people to understand:
At Via, Change Management is the underlying flavour of everything we do, especially as we support organisations beyond software selection and configuration.
We help coordinate stakeholders, guide implementation, clarify roles, manage vendors, test workflows and support adoption. Digital transformation does not end when the technology goes live, that’s when the Change Management work done (or skipped) throughout starts becoming visible.
Sources: BCG, Agents Accelerate the Next Wave of AI Value Creation, EY, Top 10 Opportunities for Technology Companies in 2026, OpenAI, The State of Enterprise AI
The reports use different language, but their conclusions are remarkably consistent. Successful AI transformation requires:
AI does not remove the need for good Systems. If anything, it makes them even more important.
The right starting question is not: “Which AI tool should we buy?”
It is: “Which business outcome matters, how should the work operate, and what combination of people, process and technology will deliver it reliably?”
At Via, we help organisations answer those questions and turn Digital Dransformation into practical, operational change programs, and successfully embed AI in their operational Systems.
Curious to know if you’re ready for AI?
Take our free 3-minute OS Systems Maturity Scorecard and discover where your greatest opportunities for stronger systems, smarter technology and successful AI adoption lie.
