Ask most employees what they use AI for at work, and the answers cluster around the same few tasks: drafting emails, summarising meetings, generating a first cut of a document. Useful? Absolutely. Transformative? Not quite.
This is the central tension in enterprise AI right now. Organisations have made the investments, rolled out the licences, and run the training sessions. Yet the value being unlocked remains stubbornly surface-level. The tools can do far more than most people ask of them.
Speaking on the UC Today AI & Productivity Show, Andreas Welsch, founder and chief AI strategist at Intelligence Briefing, offered a framework that cuts to the heart of this problem. He describes three concentric circles of AI value, each one broader and more impactful than the last.
“Tools like Microsoft Copilot and other AI assistants have done a really great job at elevating the conversation and raising awareness that we can do so much more with the help of technology. But most of the time, users are limited to: let’s draft an email, let’s summarise the meeting minutes. The basic things.”
Circle one: Personal productivity
The innermost circle is personal productivity. This is where most organisations currently live. AI acts as a personal assistant: faster drafting, quicker summarisation, smarter search. The gains are real at an individual level, and for many employees this is their first meaningful experience of AI doing something genuinely useful.
Mark Nixon, AI Business Solutions go-to-market lead at Microsoft UK and Ireland, pointed to the scale of adoption underway. Microsoft reports paid Copilot seats grew more than 160% year over year, with daily active usage up tenfold. Those who engage with it are finding enough value to return.
But Welsch argues that personal productivity, while valuable, is only the starting point. For many organisations, it has become the ceiling.
He puts part of the blame on structure.
“Part of it is lack of knowledge and lack of hands-on experience. But part of it is also organisational and cultural. We need to be AI-first, we tell people to do a lot more with AI, and then we leave them up to their own devices to figure out what they should do.”
Circle two: Operational excellence
The second circle moves from the individual to the team. Here, AI stops being a personal tool and starts driving process improvement. It automates the repetitive, time-consuming tasks that drain teams week after week.
“What are the repetitive tasks that you work on every week, every month, every quarter that you can automate with the help of AI?” Welsch asks. “Go broader than just how can I optimise my own day.”
This is where the practical examples start to land. Welsch shared one from a multinational manufacturing organisation. A plant manager pulled data exports from multiple systems, covering manufacturing execution, orders, finance and logistics, and tasked a research agent with finding correlations. Within a couple of hours, the team identified customers ordering the same parts multiple times a week. That opened a consolidation conversation that could cut transportation costs significantly.
Tom Arbuthnot, co-founder of Empowering Cloud and Microsoft MVP, makes a similar point.
“There’s so much capability to improve processes right now. It doesn’t have to be we’ve changed the entire business. It can just be lots of these little manual gaps that we can close.”
Nixon points to Vodafone’s RFP agents as a concrete example. Automating the manual process of compiling and responding to RFPs helped sellers handle two to three times more per week. For more on which productivity use cases are delivering results in 2026, UC Today has covered the workflows saving the most time across IT, HR, sales and operations.
Circle three: Strategic differentiation
The outermost circle is where the conversation shifts from efficiency to advantage. This is about creating things that were not previously possible: new products, new services, new offerings built on data and insights the organisation already holds.
“With the data we have, with the insights we have, with the things that make us truly unique as a company, how can we create new products, new services, new offerings?” Welsch says. “By starting with productivity, going to operational excellence, going to strategic differentiation, there’s a clear progression that is now possible.”
Nixon says commercial expectations are sharpening at this level. “Organisations have spent a lot of time with trials and pilots. What we’re really seeing now is they’ve realised this is going to add a lot of value, and they’re looking at how fast can they scale adoption, laser focused on the business outcomes it’s going to drive.”
Why organisations get stuck at circle one
If the path is clear, why do so many organisations stall? The panel pointed to three converging factors.
Time. The problem is not time to implement at an organisational level. It is time for individual employees to learn and experiment. Welsch described it from his own experience. “In my previous corporate role, I was one of the early adopters of Copilot. But quite frankly, I was so busy, my calendar was filled from morning until evening, that I had a really hard time carving out even 10 or 20 percent of a week to learn how this actually works.”
Arbuthnot explains: “It’s the classic sharpening the saw problem. I know this tool can help me, but I haven’t got time to sharpen it.” Organisations need to do more than hand out licences. They need to create space for experimentation and normalise early failures.
“I’ve just burnt three hours trying to automate a 30-minute process and I didn’t succeed. That being internally okay, because you were trying to be more productive, is very easy to say. But making that room for people in the enterprise is going to be really important.”
Habit. Nixon says the adoption challenge runs deeper than most organisations realise. “AI adoption is not just about learning a new tool. We’re talking about changing deeply embedded work habits. These are hardwired habits built up over years and years. A one-hour training course is not going to do that.” Sustained change programmes, with champions, continuous learning and workflow-specific support, are what actually shift behaviour.
Measurement. Without clear metrics beyond basic usage, who logged in and how many prompts were submitted, organisations struggle to show the value they generate. That makes it harder to justify further investment or build the internal case for moving to circles two and three.
Making the progression real
The good news is that the tools enabling circle one are also the gateway to circles two and three. No new platform is needed, just a new way of thinking about what those platforms are for.
Nixon’s advice is direct.
“Just identify one high-impact workflow and stand up a team that are close to that workflow. Don’t wait for some perfect end-to-end strategy. Pick a process where there’s some time savings or quality improvements or some measurable outcome, define what the success criteria is, and give the team 90 days.”
Welsch’s closing point is straightforward. The framework is not about leapfrogging to strategic differentiation overnight. It is about understanding where you are, being honest about the gap, and building the internal capability to move forward.
Most organisations are in circle one. The tools to go further are already in their hands.
This feature is based on the UC Today AI & Productivity Show, Episode 1.
