Maybe we don't need the PDF after all

"Can you make a PDF of this chart?"

The question came from an accountant who uses Capassa. She was heading into a client meeting and wanted to bring along a chart showing how the company had developed. A completely natural request, and the easy answer would have been simple: Yes, of course we can make a PDF.

But before we did, I asked: "What are you going to use it for?"

She wanted to show the chart to the client and use it as a starting point for a conversation about how the company was doing. That's when I showed her a different option. "Try right-clicking on the chart."

She did. And instead of just getting an image of the trend, the system produced the entire analysis for her. She still got the chart and the historical development she wanted to show the client. But she also got a full analysis of what the numbers showed, why the trend looked the way it did, what she should pay attention to, and suggestions for what to bring up and discuss further with the client.

It all started with a right-click.

Her reaction came fast: "I don't even need the PDF now. This is so much smarter."

That small exchange stuck with me. Not because the accountant had been thinking wrong. Quite the opposite. She did exactly what most of us do when we gain access to new technology: we start from the way we already work. She had a chart, she was heading into a client meeting, and she wanted to bring the chart.

But in the meantime, the technology had changed what was possible. And that raises a question far bigger than one PDF:

If we were solving this task from scratch today, with the technology we now have available, would we still do it the same way?

From one chart to a different way of working

The accounting industry has long talked about accountants taking on a bigger role as advisors. That makes sense. The accountant knows the client's finances, sits close to the numbers, and has a starting point few other advisors have. Even so, it isn't necessarily easy to turn that ambition into part of the working day. When we talk to accounting firms about advisory work, a few challenges keep coming up:

1. Where is the time supposed to come from?
2. How are employees supposed to do the actual analysis work?
3. How do more people become confident enough to take on an advisory role with the client?

This is where the story about the chart gets interesting. The accountant didn't have to learn AI. She didn't have to write an advanced prompt. She didn't have to set up an agent or understand the technology behind it. She right-clicked. The analysis work that would otherwise have required her to go through the numbers herself, assess the trend, look for patterns, and prepare what to bring up with the client could now be done automatically.

That frees up time. But maybe even more importantly: it changes what she can spend her time on. Instead of spending time producing the analysis, she can spend it understanding the analysis, weighing it against what she knows about the client, asking the right questions, and discussing what the client should do next.

And here's the key: the technology doesn't take over the advisory role. It can take over parts of the work that has to happen before good advisory work can begin.

When efficiency gains are no longer enough

This is exactly where a fresh analysis from Boston Consulting Group gets interesting. BCG's Applied AI Index 2026 draws on responses from more than 1,300 executives and senior leaders across more than 20 industries. Almost half of the companies in the survey now create value with AI. But only 7.5 percent fall into the group BCG calls "future-built," the most AI-mature companies. These deliver 2.3 times the shareholder return and 2.8 times the EBITDA growth of those furthest behind.

What are they doing differently? One of the most interesting findings is that the answer isn't primarily about more technology. BCG describes AI transformation as an organizational change, not just a technology upgrade. Their well-known 10-20-70 rule says the same thing: 10 percent of the effort is about algorithms, 20 percent about technology and data, and 70 percent about people, organization, and processes.

That shifts the question. It's no longer just: which AI tools should we adopt?

A far more interesting question becomes: what does the technology make possible to do differently?

BCG finds that the most mature companies concentrate their AI effort on fewer, important work processes and work with entire workflows.

And that brings us back to the PDF. We at Capassa could have built a PDF button that made an existing task a bit easier. But the way of working would, in principle, have stayed the same. When the analysis can instead be done automatically, something else happens. The accountant can spend less time producing the groundwork for advisory work and more of her expertise on understanding the client, assessing the context, asking questions, and discussing which decisions should be made.

It isn't just efficiency. It's a different division of labor between human and technology.

The accountant doesn't need to become an AI expert

When we talk about AI and advisory work, it's easy to set the bar unnecessarily high. It can sound as though the accountant now has to learn prompting, understand agentic AI, build their own agents, and keep up with a technology that changes week to week. I think that's the wrong starting point. The accountant should first and foremost be good at what the accountant already knows how to do: understand the finances, know the client, see the business in context, and use their professional judgment to help the client make better decisions.

The complexity should live in the technology, not with the accountant.

The story about the chart illustrates this well. The accountant wrote no prompt. She built no agent. She right-clicked. Behind that one right-click, the technology could do the analysis work. But that doesn't mean the accountant should uncritically pass along whatever the system suggests.

Quite the opposite. That's exactly when her professional judgment becomes even more important. She knows the client. She knows what's happened in the business. She can weigh the analysis against reality, ask follow-up questions, and discuss with the client which actions actually make sense.

AI can do the analysis work. The accountant brings the context, the judgment, and the human conversation.

BCG's survey also finds that AI is largely expected to change work rather than simply eliminate it. 89 percent of the companies in the survey expect AI to create new work, while 11 percent expect the technology to primarily replace existing jobs. Among the most AI-mature companies, 55 percent are already working strategically on which roles should exist in a more agent-based world, compared with 17 percent among the companies furthest behind.

Maybe, then, the most interesting question isn't whether AI will replace the accountant, but rather: what can the accountant spend their expertise on once the technology can handle a larger share of the analysis work?

From a few advisors to an advisory organization

This gets even more interesting once we lift our gaze from the individual accountant to the entire firm. For many firms, the challenge isn't that nobody can do advisory work. Often there are already employees who are very good at it. The challenge is doing it systematically across employees and clients.

If every accountant has to figure out on their own which numbers to analyze, how to analyze them, which patterns matter, and how to turn the analysis into a good client conversation, advisory work becomes both time-consuming and dependent on the individual. That also makes it hard to scale.

But if the analysis work can be automated and carried out using a standardized methodology across clients, the starting point changes. More employees can meet the client with a structured picture of what has happened, which conditions deserve a closer look, and which issues might be worth discussing.

And at the same time, something else happens that I think matters: the technology can become part of the learning. When the accountant is shown, again and again, which financial patterns matter, what she should pay attention to, and which actions might be relevant, she also gets a professional foundation to learn from. Analysis by analysis. Client by client.

That starts to address several of the barriers accounting firms themselves point to when we talk to them about advisory work:

Time. Tools. Expertise. Confidence.

Time is freed up because the employee doesn't have to do the entire analysis manually before the client meeting. The tool does more than present numbers. It contributes the analytical groundwork itself. Expertise can develop because the employee gets to see which patterns matter and what's worth investigating. And confidence can grow because you don't walk into the advisory conversation with a blank page. You have a professional starting point to assess, challenge, and build the conversation around.

Over time, this can have an effect bigger than the time savings alone. The firm can build a shared methodology for advisory work. Clients can get more consistent, more professional follow-up, while more employees develop advisory skills through their everyday work. At that point, it's no longer just about making one accountant more efficient.

It's about the ability to make advisory work more systematic, more standardized, and more scalable as a service, using the people the firm already has.

Maybe we don't need the PDF

It started with a simple question: "Can you make a PDF of this chart?"

We could, of course, have done that. But sometimes the most valuable thing we can do with new technology isn't to build what we already know, just a bit faster and a bit easier. Sometimes we should instead stop and ask: what are we actually trying to achieve?

For the accountant, the goal was never to produce a PDF. The goal was to walk into the client meeting better prepared, understand what the numbers were saying, and be able to help the client figure out what to do next. When the technology can do the analysis work, it doesn't just change how long the task takes, it changes what the human spends their time on.

And maybe that's exactly where the biggest opportunity lies. Not in replacing the accountant with AI. Not in turning every accountant into an AI expert. And not in piling yet another tool on top of an already busy working day. But in letting the technology do more of what the technology is good at, so people can spend more time on what they're good at.

That's also why BCG's analysis is interesting. The most mature companies don't treat AI as an isolated technology project. They concentrate their effort on important work processes, redesign workflows, and invest in people, organization, and governance at the same time.

So the big question for the accounting industry may not be: how can we use AI to make what we already do more efficient?

But rather: if we were designing this work process today, with the technology we now have available, how would we go about it?

Sometimes the answer can start with something as simple as a right-click. And sometimes we discover we didn't need the PDF after all.

Source: BCG, The Formula for Winning Enterprise AI Adoption in 2026 / Applied AI Index 2026