Communicated correctly, yet still not understood.
That was the topic of my masterclass at the Frankfurt Finance & Future Summit 2026. It focused largely on the role of AI, but above all on five mistakes that companies in the financial sector repeatedly make in their external communications. It makes little difference whether it’s a large asset manager or a start-up: The same patterns always emerge – regardless of whether the budget is large or small. Whilst the processes vary from company to company, the problems are similar. That was what this masterclass was meant to be about, and I found it fascinating, given the small group of participants, just how tense and focused the atmosphere was.
But what was it actually about? Not about texts, nor about the quality of texts in the age of AI. In fact, we delved a level deeper. It was about organisational matters and processes. AI is a huge topic everywhere at the moment, but let me say this up front: AI can help prevent these mistakes, but it can also exacerbate them. I opened the talk with five questions, and I’d like to encourage you, as you read this blog post, to work through the questions yourself and answer each one with a ‘yes’ or a ‘no’.
- Have the three most knowledgeable people in your organisation published anything under their own names this month?
- Have you personally checked this month what comparison websites and AI searches are saying about your most important product?
- Could you issue a statement by 2 pm today regarding a news item from this morning, with all the necessary approvals in place?
- If the key figure behind your communications strategy were to leave the company tomorrow, would the strategy continue unchanged?
- Do you know which of your public posts in the last quarter sparked discussions with customers?
My theory is that anyone who answers ‘no’ to even one of these questions is likely to be making one of the five mistakes I discussed in my talk and in the masterclass.
The outcome of the vote
Thirteen participants voted anonymously. This is not representative. Unfortunately, the results could not be displayed live in the conference room (due to technical issues). But the results are now in. They are revealing.
- 1. The best minds have published under their own names (over the last four weeks or so)
- Yes: 7
- No: 5
- Don’t know: 1
- 2. Portals and AI searches checked personally
- Yes: 9
- No: 3
- Don’t know: 1
- 3. Comments by 2 pm
- Yes: 6
- No: 5
- Don’t know: 2
- 4. The strategy is continuing without its key figure
- Yes: 5
- No: 8
- Don’t know: 0
- 5. Knowing which posts trigger customer discussions
- Yes: 2
- No: 9
- Don’t know: 2
So let’s have a look at that.
- Of the 13 who voted, only two know which of their public posts (by which I meant press releases, LinkedIn posts and so on) actually triggered a conversation with a client or a potential new client most recently.
- Another interesting question was about the key figure. What happens if the key figure leaves the company? Will the communication strategy remain in place? Eight out of 13 who voted did not believe that the communication strategy would remain intact if that person left.
- The question regarding approval processes and providing a prompt response to a current market event was a bit closer: 6 ‘yes’, 5 ‘no’ and 2 ‘don’t know’ out of 13 participants. Six said the company would be able to respond to such an event itself by 2.00 pm. Five said they would not be able to, and two said they did not know.
Following the opening of the masterclass, the discussion turned to practical and procedural issues: What are we talking about now, and where do the real problems lie? Most people wanted to talk about working with their C-level executives and collaborating with the compliance department, as well as how one actually measures success in communications work within the financial sector.
The five mistakes:
1. The best senders are left to their own devices
After all, the financial sector is home to an incredible number of clever people in various specialist roles. Yet some of them are completely overlooked by the general public. Companies regularly publish analyses and market commentary, but often these aren’t really written in their own voice; instead, they’re very polished. This is particularly true at C-level. Many board members don’t want any help at all and say: “I’ll do that myself”, or: “I don’t have time for that.”
There are many reasons for this. Some of them relate to liability. So people prefer to hide behind compliance and the organisation. Yet editorial teams and the public are often far more interested in the views of competent individuals. The problem is that, within the communications team or even at the agency, often nobody feels responsible for really putting these people – the good ones – in the spotlight.
We took a completely different approach with a financial services provider, building it up strategically from the very start. We began before we’d even written any copy. For each member of the executive board, we clarified a number of points.
So: What do you want to achieve? Who are your role models? Where can we look to see what these role models are writing, for example on LinkedIn? We looked at this closely and then analysed it in detail. We analysed the board members’ own posts that they’d published to date. And we analysed what their role models – or even their anti-role models – had published. Then we looked at the patterns of reaction, the networks involved and the feedback. We analysed all of this thoroughly and then derived topics from it and produced initial drafts. The result is that C-level executives, or board members, never find themselves facing a blank page. I put it this way: the person edits their own voice.
AI can be a huge help here, and – something that’s often misunderstood – it can, above all, help strike the right tone, because it can be trained to do so. But without a human being to edit it afterwards and make it a bit more human – and perhaps a bit more imperfect and subjective – you simply end up with that uniformity, that ‘AI slop’ we’re all familiar with. I summed it up in my talk as follows: companies communicate facts; people give them meaning.
2. I’ve produced a lot; others will tell the rest of the story
Many companies in the financial sector invest an incredible amount of money, energy and time in producing content. But then there’s an incredible amount of content circulating out there on the internet that you have no overview of. Comparison websites write about you. Particularly in the brokerage sector and where it’s more B2C-focused, there are also affiliates who write, or so-called creators, and other partners in the field of sales support. They all write about your brand, your company and your products. But do companies actually know exactly what they’re writing? It becomes particularly tricky because AI search engines then use precisely these sources to generate summaries, which are then presented to users of AI search engines as fact. Yes, and that’s precisely what sets this apart from the working methods that have existed up to now. There are people responsible for everything – there are those in charge of PR, and those in charge of paid reach. Comparison sites, affiliates and other business partners often fall through the cracks these days.
So in our day-to-day work, we regularly come across incorrect information circulating online about our clients and their products, and we then contact the relevant website operators with the correct details. Sometimes product features are incorrect; sometimes the information is out of date. Sometimes a publisher has even written the company profile themselves. That can be fine, of course. But it happens very regularly. What’s interesting here is that whoever commissions such partners may, under certain circumstances, also be jointly liable for their statements, provided they are promotional in nature (Section 8(2) of the German Unfair Competition Act (UWG)).
You can’t really stop third parties from doing anything. But you can keep an eye on things. You can provide them with up-to-date facts. You can regularly check what they publish. But the best approach is to take a truly proactive stance. And that means supplying up-to-date, accurate content about your own company, products and services directly to the platforms where it will be published. In our line of work, publishing regular market analyses is a great way to position oneself. Ideally, you should have direct control over what is published – ensuring it is accurate and compliant – as this means you’ll have less to chase up and correct later on.
3. When it comes to approvals, everything takes the long way round
The third mistake concerns timing. Anyone working in communications within the financial sector will certainly be familiar with this. The market commentary is ready. It then goes through compliance, and by the time it’s approved, the market has long since moved on. We need solutions to enable us to communicate effectively. It becomes problematic when a policy issue goes through exactly the same approval process as a current market commentary. A response to a client’s enquiry, for example, is a completely different matter. We must distinguish between urgency and risk.
Establishing this distinction, defining it once and embedding it in our processes naturally takes work. But in the end, it’s less work than constantly putting everything through the most stringent checks. Incidentally, that doesn’t mean I’m in any way against compliance and checks. That’s important to me. Where there are guidelines requiring mandatory approvals before publication, that must remain the case. I’m not advocating for less control. I want control to be in the right place. Regulators, laws and regulations require appropriate controls, but they do not necessarily require an identical, maximum-level approval process for every single publication, no matter how small, or for every little thing a company does.
I have not only witnessed but also helped shape how this works very differently – and better – at a very large global consumer goods company. We defined processes there. We drew up a handbook setting out very precise processes. What did it say? A response is sent out only if there is an approved template for it. If there isn’t one, then the text is drafted, agreed upon and fine-tuned for future use. We always classified this, as well as current events, according to three criteria: relevance, reach and dynamism. Can this be applied to a company in the financial sector? Perhaps not exactly as it stands. But I do think the logic applies: it’s best to approve things in advance that can be said, rather than waiting until the very end. Incidentally, this is also a huge advantage for procedural work with AI: what can be released, and what cannot.
Ultimately, in my consultancy work, I often advocate a two-track approach. One of these tracks is the ‘fast lane’. It has defined response times, and the fast lane also features a written mandate and a final check where this is simply necessary based on the level of risk. And with processes like these, AI can actually make the fast track really quick. If, on the other hand, you don’t introduce such processes, then the AI simply produces more drafts for the same queue – that frustrating queue we’re already familiar with.
4. The strategy depends on individuals
Well, when they see the subheading ‘Strategy depends on people’, many might think, ‘That’s exactly how it should be.’ The communication strategy should indeed depend on people. Yes, that’s certainly true. But it can also be problematic if you’ve gone through a very defined, very strict strategic process and ended up with a very precisely defined strategy. That’s because it can sometimes start to erode. For example, due to staff changes: a key corporate influencer leaves or someone new joins the team. The new person may not have been involved in defining the strategy. That’s when the friction starts. With every publication, the strategy is challenged. There’s a spanner in the works. When in doubt, you end up defending or re-arguing the strategy all over again every single time. I’ve experienced this myself. It slows everything down and is incredibly frustrating.
The question is, how can we do things differently now? I believe the answer is that strategy is never ‘finished’. Often, a strategy is treated like a document. At some point, the document is finalised, approved, and then filed away. Individual people then take on the task of interpreting and implementing the strategy. If you ask yourself right now: ‘It’s 10 pm and something happens – who’s authorised to respond?’ Is it a single person, or is it a role within your organisation – one comprising an experienced team – that’s authorised to respond? If only one person’s name springs to mind, you’re right in the midst of Mistake 4.
So if the strategy depends on individuals, then it is no longer a strategy. Even AI won’t help here. Because if AI is the gatekeeper and monitors adherence to the strategy (and, incidentally, it’s really good at that), then we end up with major contradictions between the implementation of the strategy and what individual people actually do. This, in turn, ties in with Mistake 1. Individuals aren’t supported in their external communications – in other words, it’s the organisation that speaks to the outside world. It should actually be the other way round. Outwardly, it’s the names, the board members and the brand ambassadors; whilst internally, in the coordination and execution of the strategy, it’s the roles that matter, not the individuals.
5. Everyone claims success, but no one measures the impact
Writing reports – or ‘reporting’, as it’s known – has been part of my day-to-day work for many, many years. And I see lots of different disciplines producing their own reports: public relations, online marketing, search engine optimisation. Most recently, GEO has been added to the list – search engine optimisation for AI engines. Then there are reports from social media teams, reports from trade fairs – reports galore. We always count what’s easy to count. You can usually count things straight away, but where’s the impact, and where’s the value added to the business?
It all starts with the impact. The impact often only becomes apparent later. It often doesn’t even show up in the tools at all. On top of that, AI trivialises metrics such as reach, interaction or engagement. And by that I really mean trivialises them. Such figures actually say less and less. In fact, we communicators distinguish between four levels. Level 1 is effort, Level 2 is reach, Level 3 is impact, and the highest level is value contribution. I regularly find that measurement actually stops at Level 2 – at reach – even though everyone supposedly uses it to report on impact.
I always think it’s ideal when we manage to produce comprehensive, interdisciplinary reporting that spans all communication disciplines. That’s the ideal scenario and, in my view, it’s achievable. But there’s a long way to go. You can also start on a smaller scale. It doesn’t even cost anything. I urge you to do this: make a note of who from your target audience has reacted to a post? In other words, who? Not how many, but who, and what position does each person hold? What value could it bring to the company to forge closer ties with this person in their role – whether as a customer or perhaps as a future employee?
What remains human?
The masterclass at the Frankfurt Finance & Future Summit was, of course, promoted through various announcements. One announcement focused on the question of what actually defines human communication today. What defines it, given that AI is producing ever more content? We’ve now looked at these five mistakes. Having considered them, I’d say: it’s important to show your face, to decide what goes out and what doesn’t, and to take responsibility for it, based on strategy, conviction, measurement and common sense.
This is not merely a theoretical point from a legal perspective either. In May 2026, the Higher Regional Court of Hamm ruled that a chatbot’s responses constituted a business act on the part of the operator. Admittedly, the case involved a clinic rather than a bank, but the argument that ‘the machine said it’ did not help the operator. The machine writes. A human being must take responsibility for it.
This brings me back to the sentence with which the organiser announced my masterclass: ‘If you want to shape the future, you must first be understood.’
