Are Your Job Adverts Already AI-Ready?

The title is still the same. But the job underneath has long since become a completely different one.

We have several job adverts out again at the moment. New colleagues are joining, roles are being refilled, and of course we keep asking ourselves the same question: who are we actually looking for? What does a person need to be able to do today? And what will this person concretely do for us later on?

It is a perfectly normal situation. Until you open a job description that may be two years old.

Then you read phrases such as “researching potential customers”, “maintaining the customer database”, “creating presentations”, “summarising meetings” or “writing copy”, and you catch yourself thinking: do we still actually do it that way?

Six months ago, perhaps. Today, in part, hardly at all.

Not because these tasks have disappeared altogether. But because the way we carry them out has changed within a very short time. Research now often starts with AI. Nobody has to summarise a meeting manually for 45 minutes any more. A presentation does not necessarily start on a blank slide. A sales employee no longer has to gather information from five different sources. A first draft of a text no longer takes two hours. From CRM data, summaries, next steps or warning signals can be derived automatically.

And sometimes you really do sit there and wonder: How did we ever manage all of this before?

Of course we managed. With more clicks, more copy & paste, more Excel, more follow-up questions, more searching, and probably with a surprising amount of working time whose waste we simply took for granted.

That is exactly why I think one of the most discussed questions at the moment is a little too simple.

In short: At the end of this article you will find a simple method for assigning every task in your job descriptions to one of four categories:

• Human: needs experience, empathy or judgement
• Human plus AI: stays with the human, but becomes faster with AI support
• Automate: a system can take it over independently
• Abolish: you would not introduce it today

This lets you write job adverts that match what your team really does, rather than what it did two years ago.

Which jobs will AI replace?

Perhaps that is not the decisive question at all.

A job rarely consists of exactly one activity. A job is usually a bundle of ten, twenty or fifty different tasks that were at some point grouped under one job title.

A sales employee researches customers, finds contacts, prepares conversations, maintains a CRM, writes emails, creates quotes, carries out follow-ups, advises customers, negotiates and, in the end, tries to close a deal.

We call all of that “sales” together.

Yet these are entirely different activities. Some require an understanding of people. Others analysis. Others data entry. Others organisation. Some boil down to little more than moving information from A to B.

And it is precisely this bundle that artificial intelligence is now starting to take apart.

An agent can prepare the research. Conversation details can be documented automatically. CRM entries can be generated from emails or meetings. Quotes can be prepared on the basis of existing information. Follow-ups can be suggested. A system can recognise that a customer has not responded for three weeks. Perhaps one day an agent will even decide that now is the right moment for a new contact.

Does that mean there will be no sales left? Of course sales will remain.

But perhaps the job will increasingly consist of the things we originally needed humans for: understanding, listening, putting things in context, building trust, deciding and negotiating.

The job does not necessarily disappear. It changes its composition.

That is why the more interesting question is not: Which jobs will AI replace? But rather: Which tasks within a job will still exist in three or five years?

How does AI change job descriptions?

This is exactly where classic job descriptions start to wobble.

Many job adverts mainly describe activities. But it is precisely these activities that are changing fastest right now.

In my view, companies should therefore define roles more through outcomes and responsibility and less through individual manual work steps.

Not “You maintain our CRM.” But: “You make sure our sales team can access up-to-date and reliable customer information at all times.”

Not “You create reports.” But: “You make sure management recognises early where action is needed.”

Not “You research potential customers.” But: “You identify relevant sales opportunities and prepare them so that concrete next steps follow.”

And not “You write training materials.” But: “You make sure people understand new knowledge and can then apply it.”

The outcome stays the same. The way to get there is allowed to change. And that is exactly what makes a job description much more durable.

Which tasks can already be supported by AI today?

For business owners, this is a far more practical question. The answer: a surprising number by now.

Research, summaries, meeting follow-up, first data analyses, draft texts, quote preparation, document comparison, minutes, CRM summaries, email drafts, the evaluation of customer feedback, knowledge search, project preparation or the identification of open tasks can already be supported in a useful way today.

Not every activity can be fully automated. And not every one should run without human oversight.

But for a surprising number of tasks, the question is no longer whether AI can help, but how the process should be sensibly designed.

One mistake should be avoided here, though, one that companies already made in earlier waves of digitalisation: simply making a bad process faster with new technology.

A bad process remains a bad process. Only faster.

Which tasks should a company automate with AI, and which should it abolish instead?

In my view, this is one of the most exciting questions of all. Some activities exist only because our systems used to be unable to do any better.

We transfer information from an email into a CRM. We turn meetings into minutes. We copy data from one Excel file to another. We gather documents. We send status requests. We check whether someone has updated the data.

For decades this was normal office work. Today, for tasks like these, you should ask yourself two questions.

First: Why is a human still doing this? And straight after: Why are we doing this task at all?

Perhaps the process does not need to be automated. Perhaps it should simply disappear. Anyone who wants to automate processes systematically should therefore always ask this question first, before choosing the technology.

What does AI mean for sales, project management and marketing?

Let us take a concrete example from our own job adverts. For a sales internship, it could easily have said in the past: researching potential customers, maintaining the customer database, creating quotes and following up sales activities.

A few years ago that was completely normal. Today I find a different sentence more interesting: Manual data maintenance is not a learning goal with us.

Of course an employee has to understand what happens in a CRM. But if someone spends two hours every morning copying information from other sources, I would now rather expect the question to arise at some point: can this not be automated?

After all, the more valuable task is not entering data. It lies in recognising a sales opportunity, sizing up a company, understanding a customer and deciding what should happen next.

It is similar in project management. I do not necessarily need someone whose greatest strength is writing particularly beautiful minutes. I need someone who understands after a meeting which decision is missing, where a risk is emerging, which requirement has not been properly understood and who is currently waiting for whom.

The machine is welcome to write the minutes. Someone should still understand the project.

In marketing, too, the skill is shifting. “Being able to write” was long a clearly defined ability. Today practically anyone can produce a usable first draft in a few seconds.

That does not mean writing skill becomes unimportant. On the contrary. The questions become more important: is the topic interesting at all? Is the claim correct? Is it boring? Does it suit the target group? Is the perspective new? What story is in it? Is the text merely linguistically clean, or genuinely good?

A good marketing employee may have to spend less time getting words onto an empty page. In return, they must be better at judging which words belong on that page in the first place.

Does AI make jobs accessible to more applicants?

This is where the discussion gets particularly interesting for me. Because AI and modern software change not only tasks. They may also change the question of who is suited to a job in the first place.

We have experienced this ourselves. Some time ago, at one of our locations in Germany, a woman applied who originally came from accounting and had only a few years left until retirement.

If you had looked only at the CV, the answer would probably have been: not a fit. Her professional profile did not really match what we had originally been looking for.

But she stood out for something else. She was extremely willing to learn, reliable, attentive and pleasant to communicate with. Above all, she was not afraid to say: I don't know that yet. Please show me.

So we tried something. We asked less: Can she already do everything in the job description today? And more: Can she learn to work with the tools she needs for this role?

She received clear onboarding, defined processes, the relevant systems and a fixed framework for working with AI.

This framework is important. AI in a company must not mean that every employee copies information indiscriminately into any system. Rules are needed. Which data may be used? Which information does not belong in a public system? When must an answer be checked? When is a human absolutely required?

Within this framework, however, something interesting happens. An employee does not have to collect every question of understanding and wait until an experienced colleague eventually has time for it. They can have a term explained, have a complicated relationship simplified, think through a first approach to a solution or check what information is still missing.

AI thereby also becomes a kind of learning and orientation system that is available at any time.

An applicant whose CV did not really fit at first became someone who, with the right tools, could reliably find her way into a new working world. And I find that much more interesting than the question of whether someone did exactly the same job ten years ago.

A project manager the perfect CV would probably never have found

A second example is one of our project managers. Today he is among the colleagues who are particularly popular with customers, in Germany as well as with Austrian customers, which can be quite an achievement given partly different expectations and communication styles.

His original background? Advising customers in an electronics shop.

Not a classic career path for a project manager in digitalisation projects. No ten-year CV from consultancies. No perfectly fitting title.

What he did bring, however, was something much harder to teach. He could deal with people. He listened. He tried to understand what a customer really wanted. He stayed friendly when things got complicated. And he thought in terms of solutions.

The rest can be structured. Project methods can be learned. Software can be learned. CRM systems can be learned. Documentation standards can be learned. With the right tools, research, preparation and documentation can also be done much faster today than before.

Someone with a background that looked unusual on paper thereby became a very good project manager. Not despite the CV not fitting perfectly. Perhaps partly because of it. Customer orientation and solution thinking were there from the start. The tools came afterwards.

Both stories also reflect our experience from over a thousand projects: professional perfection on paper was rarely the decisive success factor.

Are companies still looking too much for ready-made people?

This leads to a thesis that is certainly open to debate: Many companies still look for people who already fit an existing position exactly. AI could increasingly turn this principle on its head.

Perhaps we should look less for fully finished candidates and more often for people who communicate well, are reliable, take responsibility, want to learn, understand problems, try out new tools and know when they need support.

Because much of what used to count as essential operational skill can now be learned faster or supported technically. That does not mean expertise becomes unimportant. But perhaps the balance shifts.

Do generalists become more important through AI?

My answer would be: good generalists, probably yes. I like to call this the Swiss Army knife principle.

A Swiss Army knife does not replace a professional workshop. But with the right toolbox, a good person can solve a surprising number of problems. AI, automation and modern software are currently expanding this toolbox considerably.

An employee may no longer need to research perfectly, write perfectly, present perfectly, know every piece of software by heart, build every report themselves and start every document from scratch, all at the same time.

Instead, they need to understand: Which tool helps me with which problem? And then: Can I judge whether the result is good? That is a different competence profile.

Perhaps we are therefore experiencing a small renaissance of the generalist. Not of the person who can do a little of everything and nothing properly. But of a person who understands problems, can connect different disciplines and knows when a specialist is additionally needed.

This could also change recruiting. Someone is an excellent project manager but not a good designer? Perhaps not a big problem any more today. Someone understands customers extremely well but is slow at writing? Also solvable. Someone has a good analytical understanding but is no Excel wizard? Possibly less critical than before.

Does AI make recruiting easier?

In part, perhaps genuinely so. Many job descriptions today are astonishing wish lists. They look for someone with experience in marketing, project management, analysis, presentations, CRM, research, Excel, copywriting, data, automation and communication. Preferably all at once, and of course with five years of professional experience.

Perhaps this person exists. But rarely.

If, however, a candidate brings eighty per cent of the truly important skills and can quickly build up missing operational skills with the help of modern tools, they could in the long run be the better candidate than someone who masters a particular system perfectly today but rejects change on principle.

Perhaps our idea of the perfect applicant is therefore changing too. One of the most important qualifications in future may be: This person can acquire the missing ten per cent. And six months later the next ten.

Does AI lower the barrier to entry for knowledge work?

For certain activities, probably yes. But that is only half the truth. Because while the operational barrier to entry may fall, the demands on judgement rise.

An employee has to recognise: is this answer correct? Is information missing? Is the source credible? May I use this data? Is this result appropriate for the customer? Should this decision be prepared by an AI at all?

That is why I would never claim: with AI, everyone can now do everything. That would be nonsense.

My thesis would rather be: With AI, significantly more people can take on tasks that used to require a much narrower profile. At the same time, it becomes more important to be able to assess results and take responsibility.

In short: The barrier to entry falls. Responsibility rises.

Can AI replace missing expertise?

No. An AI model does not turn a person without medical training into a doctor. It does not turn a career starter into an experienced management consultant. And ChatGPT does not make a person without judgement a good project manager.

But between “I am an expert” and “I can't do that” there is an enormous area. That is exactly where AI is changing a great deal right now.

People can get up to speed faster, have terms explained, research connections, compare alternatives, prepare calculations and structure information.

Perhaps, then, not everyone becomes replaceable through AI. Perhaps every good employee first becomes more versatile.

Which skills become more important in the age of AI?

When routine knowledge becomes more readily available, other skills rise in value.

Judgement becomes more important: is the result really correct? Understanding of context becomes more important: does the solution fit the actual problem at all? Communication becomes more important: do I understand what the customer or colleague really needs? The ability to learn becomes more important: can I put a new tool to productive use quickly? Problem solving becomes more important: can I break a large problem down sensibly? And curiosity becomes more important: when doing recurring work, do I even ask whether there is a better way?

Interestingly, some of these are very old human skills. AI does not make them less important. Perhaps it even makes them more valuable.

What happens when every employee gets their own AI agent?

This is where it gets really exciting. Today many people still use AI on the principle of: ask a question, get an answer, process the result further.

But development is increasingly moving towards systems that can work through longer chains of tasks themselves. So let us imagine a sales employee who at some point directs a research agent, a CRM agent, a follow-up agent and an agent for quote preparation.

Their work then consists less of carrying out every single step themselves. Instead, they have to define goals, set priorities, check results, handle exceptions and make decisions.

That sounds surprisingly like leadership. Except that part of the team consists of software.

Will AI agents eventually need their own roles in the organisation chart?

Perhaps. At the very least, companies should allow the question.

If an agent independently processes customer information, carries out tasks or prepares decisions, companies must define what it is allowed to do, which data it can see, when a human has to approve and who is responsible when errors occur.

Perhaps companies will one day define not only employee roles but also agent roles. Not necessarily with a photo on the team page. But structurally, very much so.

Can AI also relieve managing directors and managers?

Absolutely. Perhaps one should even start there.

Because a surprising number of managing directors spend a considerable part of their time on searching for information, status queries, meetings, approvals, reporting, follow-up questions, emails and coordination.

If a company is to become less dependent on its owner, one question is therefore particularly interesting: Which of the managing director's tasks actually have to be done by the managing director?

And which ended up there only because information, processes and responsibilities in the company are not properly organised?

Perhaps the most important job description a business owner should revise in the age of AI is therefore their own.

What does a modern job description look like in the age of AI?

A modern job description should describe less how each individual task is carried out today, and explain more why the role exists, what outcome it is to deliver and what responsibility a human bears.

It should therefore answer questions such as: why does this role exist? What problem does it solve? What outcome should result? How will we recognise success? Which skills must actually be present? Which decisions stay with the human? Which tools are available? Which tasks are allowed to change? And what should deliberately remain human?

A good job description therefore says less: This is how you do your work. And more: This is what you are responsible for.

How do I tell whether my job description is still up to date?

You do not need a big workshop for this. Take an existing job description and check every single task against four categories.

• Human: This task needs experience, responsibility, empathy, creativity or judgement.
• Human plus AI: A human remains responsible, but can do the task considerably faster or better with AI.
• Automate: A system should take over this task as independently as possible.
• Abolish: If we built the process from scratch today, we would not introduce this activity at all.

This simple exercise can show surprisingly quickly just how “pre-AI” a job description actually is.

Which questions should managing directors ask their employees now?

Instead of only asking “Where do you see yourself in five years?” in the next appraisal meeting, I would try three other questions.

• Which three tasks do you still do manually today that you believe we will do differently in twelve months?
• Which activity regularly costs time even though it creates hardly any value for customers or the company?
• And what would you like to do more often if routine work were taken off your hands?

These three questions can already yield a surprisingly good automation roadmap.

Does a company now have to rewrite all its job descriptions because of AI?

No. But companies should start to question them regularly. For highly digitalised roles, perhaps every six or twelve months.

Not with the aim of constantly producing new documents. But with the simple question: Has the work changed? Because by now it sometimes changes faster than the official role.

How do you recognise an applicant who is ready for AI?

I would not primarily look for “ChatGPT skills”. At some point that will probably be about as meaningful as “can use Google”.

More interesting are questions such as: how do you approach a problem you have never solved before? Which recurring task did you most recently automate? How do you check whether an AI answer is right? When would you deliberately not use AI? How do you get to grips with a new tool?

That is more likely to find people who cope with a working world in which the toolbox gets bigger every few months.

Would you still build your company the same way today?

For me, this is the most exciting question in the end. Imagine you could rebuild your company completely from scratch tomorrow.

No existing software. No old processes. No Excel files. No historical responsibilities. No job descriptions. No “We have always done it this way”.

You have your customers, your knowledge, your employees, and in addition all the technologies available today.

Would you really organise the company the same way again?

Would you create the same positions? Distribute the same tasks? Build the same departments? Or would you first ask: what does a human really need to be good at? What can software handle? What can AI prepare? What can agents take over independently? What can be automated? And what work should we stop doing altogether?

I suspect that this is where the real change brought by AI lies. Not in whole professions suddenly disappearing on a Monday morning. But in something much less spectacular.

A piece of research here. A piece of documentation there. An evaluation. A bit of data maintenance. A preparation. A routine decision. Task by task.

Until at some point we look at a job description and realise: the title is still the same. But the job underneath has long since become a completely different one.

Perhaps it is time to treat it that way. If you do not want to take this step alone, you can tackle the topic of rethinking your business model and organisation digitally together with us.

---

Frequently asked questions about AI, jobs and job descriptions

How does artificial intelligence change job descriptions?

AI mainly changes individual activities within existing jobs. Job descriptions should therefore describe fewer rigid manual tasks and more outcomes, responsibility, human skills and the sensible use of AI and automation.

Which jobs will be replaced by AI?

This cannot be answered reliably in general terms. It is more likely that many tasks within existing jobs will change first. Some activities will be automated, others will become more important, and new tasks will emerge.

Does AI make jobs accessible to more applicants?

For certain knowledge-work roles, probably yes. AI can partly bridge operational skill gaps and speed up the onboarding into adjacent tasks. Expertise, judgement and responsibility nevertheless remain decisive.

Do generalists become more important through AI?

Good generalists can benefit if they can connect different specialist areas, structure problems and use modern tools sensibly. It is also decisive to know when a specialist is additionally needed.

Can AI replace missing expertise?

No. AI can make research, preparation and onboarding much easier, but it does not completely replace sound expertise, experience and human judgement.

Which AI skills should applicants have today?

More important than knowing a particular AI tool are problem solving, the ability to learn, asking good questions, critically assessing results, checking sources and a responsible approach to data.

How often should companies review job descriptions?

For highly digitalised roles, a review every six to twelve months can make sense. In particular, you should check which tasks can by now be supported, automated or abolished altogether.

Should the use of AI be stated explicitly in job adverts?

If AI is part of the normal toolbox of the role, yes. However, the advert should ask less for a single product and more for the ability to use modern AI tools sensibly and responsibly.

Can an employee take on several areas of responsibility in future?

For certain roles, absolutely. AI and automation can help people take on adjacent activities that used to require additional specialist knowledge.

What is an AI-ready job description?

An AI-ready job description mainly defines the purpose, outcome, responsibility and necessary skills of a role. It leaves enough room for individual activities and tools to change.

Which tasks should remain with humans despite AI?

In particular, activities involving high responsibility, complex context, important personal relationships, ethical considerations or decisions with greater consequences should not be automated without checks.

How can AI relieve a managing director?

For example through automatic summaries, reporting, preparing information, meeting follow-up, research, status overviews and the preparation of decisions. Entrepreneurial and strategic responsibility remains with the managing director.

Will companies still need classic job descriptions in future?

Probably yes, but in a different form. Instead of fixing individual activities for years, roles could increasingly be defined by purpose, outcomes, responsibility and skills.

---

Change the job description, or the company behind it?

In the end, though, this discussion should not stop at HR.

If, while revising a job description, you find that employees spend a surprising amount of time on manual data maintenance, searching for information, follow-up questions, handovers, copy & paste or status queries, the real problem probably does not lie in the job description.

Then perhaps you are looking at a process problem. Or a software problem. Or an information problem. Or at an organisation that has grown historically and whose workflows nobody has fundamentally questioned for a long time.

And that is exactly where it gets interesting. Because perhaps your employee does not need a more modern job description at all. Perhaps they need a company in which they spend less time on things for which we have long had better tools.