Back to Blog

Will AI Replace Jobs? Which Will Go, Which Will Stay

Henri Jung, Co-founder at Superkind
Henri Jung

Co-founder at Superkind

A classic dark leather business briefcase standing upright, with the orange Superkind mascot standing beside it

Will AI replace jobs? Friday afternoon in a Mittelstand office in Hanover. Nadine, who has booked invoices for twelve years, reads the headline on her phone: “57 per cent of office hours automatable.” Her colleague Daniel, team lead in sales, sends her the article with a question mark. Neither wants a slogan. They want to know what happens on Monday, which parts of their work are gone next year, and what they should learn now. This guide gives the honest answer, with real numbers from the World Economic Forum, the IAB, PwC and McKinsey, and shows which tasks Superkind’s ready-made AI team can take off their desks today.

Short version: not many whole jobs go away. A lot of tasks inside most jobs change. Some tasks disappear, new ones appear and the people who stay have the time back to do work that was always more important. Superkind sits in that gap. Thomas Wolf, Leiter IT at an industrial customer, puts it this way: “Superkind doesn’t replace our employees, it complements them. And makes them better and faster.”

Key takeaways

Tasks go, not jobs - the World Economic Forum projects 170 million new jobs and 92 million displaced by 2030, a net gain of 78 million. McKinsey calls the 57 % of work hours that are technically automatable a potential, not a forecast.

Office and admin feel it first - up to 54 % of expected job transitions sit in office and administrative work. The International Labour Organization puts the share of jobs in Germany facing full automation at only 5.5 %.

What stays is judgement, trust and relationships - nursing, teaching, skilled trades, management and specialised crafts stay closest to people. The share of German jobs with high substitutability potential still rose from 15 % (2013) to 38 % (2022), so change is real.

Get a ready-made AI team for everyday work - Superkind comes with a team of AI employees (Inbox Clearer, Report Builder, Quote Writer) that run on their own and take the repetitive parts off your desk. Start for free with unlimited users, or build your own.

Keep the control, keep the data in Europe - hosted in Europe, GDPR compliant, EU AI Act aligned, no training on your data, and you decide where the AI waits for your OK.

Your AI team for everyday work

Order status for order 48213
customer@thalmann.de
Hello, can you tell me when order 48213 ships? We need it by Friday.
Inbox ClearerAI employee
New mail in service@
Checked your CRM and the shipping system. Order 48213 ships today, DHL tracking link sent. Reply drafted and sent on its own, in German.
Mails this morning41
Answered on their own28
Waiting for a person4 complaints
Routine mail runs on its own
Example: Inbox Clearer at work in Outlook

What does “AI replaces jobs” actually mean?

AI replaces tasks inside jobs more than it replaces whole jobs: an AI system does part of the work a person did before, usually the repetitive, text-heavy or data-heavy part, while the person keeps the judgement, the customer relationship and the final decision. Economists call this “exposure” or “substitutability potential” and measure it per task, not per role. One example is Superkind: a ready-made AI team that does routine mail, reports and quotes in your programs, hosted in Europe, with 1,500+ integrations into the systems where that work already lives.

The difference matters. A clerk whose day was 70 per cent typing invoices and 30 per cent calling suppliers does not lose the job if the typing goes to AI; the day becomes 20 per cent checking what AI drafted and 80 per cent calling suppliers and managing exceptions. The job description changes. The headcount sometimes shrinks over years through natural attrition, more often it stays flat while the company grows. What definitely changes is what the person spends their Monday morning on.

How it works with Superkind: your AI team for everyday work

Superkind is the AI employee for your team: you write to Superkind by email (Outlook or Gmail), in Microsoft Teams, in Slack or in the Superkind web app, like you would write to a colleague, and Superkind does the task in your programs. Hosted in Europe, GDPR compliant, with 1,500+ integrations into the systems where your work already lives: Outlook, Teams, Slack, SharePoint, HubSpot, Salesforce, DATEV, SAP Business One, Excel and DocuSign.

You do not start from scratch: Superkind comes with ready-made AI employees for the work that fills most inboxes. They run on their own in the background, and you decide where they wait for your OK. Use them right away, adapt them, or build your own.

Inbox Clearer

AI employee for the shared inbox

The service@ mailbox fills up every morning: order status, invoice copies, change-of-address, a few real complaints. As soon as a new mail arrives, the Inbox Clearer reads it, pulls the order from your CRM or the last invoice from SharePoint, drafts the answer in the right language and sends it on its own for the simple cases. Complaints and anything over a threshold wait for a person.

Works inOutlookHubSpotSharePointMicrosoft Teams+ 1,700 more
Order status for order 48213
customer@thalmann.de
Hello, can you tell me when order 48213 ships? We need it by Friday.
Inbox ClearerAI employee
New mail in service@
Checked your CRM and the shipping system. Order 48213 ships today, DHL tracking link sent. Reply drafted and sent on its own, in German.
Mails this morning41
Answered on their own28
Waiting for a person4 complaints
Routine mail runs on its own
Example: Inbox Clearer at work in Outlook

Report Builder

AI employee for weekly reports

Every Monday someone spends two hours pulling numbers from Excel, HubSpot and the DATEV export. The Report Builder does it before anyone is in the office: it opens the sources, builds the one-page report in the company template on SharePoint and posts the summary in the Teams channel. If a number drops below your threshold, it says why.

Works inMicrosoft TeamsExcelHubSpotSharePoint+ 1,700 more
Weekly reportManagement
Report BuilderAI employee
Routine · every Monday at 07:00
Weekly report ready in SharePoint. Revenue + 6.2 % week on week. New customers 11. Open quotes EUR 184,000. One alert: inbound tickets up 24 %.
RevenueEUR 318,400 (+6.2 %)
New customers11
Open quotesEUR 184,000
Runs on its own, every Monday at 07:00
Example: Report Builder at work in Microsoft Teams

Quote Writer

AI employee for quotes and orders

A request comes in by email: product, quantity, deadline. The Quote Writer finds the right contact in HubSpot, pulls prices from the Excel rate card, writes the quote in the company template and sends it via DocuSign. For quotes over your limit, management signs off first.

Works inOutlookHubSpotExcelSharePoint+ 1,700 more
Superkind
New chat
My tasks
AI employees
Quote Writer
Inbox Clearer
Report Builder
Quote for Meier Metalltechnik

For every new quote request: write the offer from the 2026 rate card and send it after my OK over EUR 10,000.

Reasoning ⌄
Quote Writer 14:22
Routine · every new quote request

Quote for Meier Metalltechnik ready. 4 line items from the 2026 rate card, total EUR 14,820 net, margin 18.4 %. Over your limit, so it waits for your OK.

Line items4
TotalEUR 14,820 net
Margin18.4 % (minimum 12 %)
Management sign-offOver EUR 10,000, so it waits for your OK.
ApproveReject
Send a message ...
Example: Quote Writer at work in Superkind

Every employee is ready to use right away. Or build your own on the platform.

Safe by default

The data that passes through routine mail, reports and quotes (customers, prices, personnel) belongs in Europe. Superkind keeps it there and never uses it to train models.

From the first task to your own app

Four stages, told once on the question of this article. Each one builds on the last:

  1. 1

    Start

    Ready in minutes. Sign up for free at app.superkind.ai, forward the first routine email from service@ to Superkind and watch it draft the answer from your CRM.

  2. 2

    Relief

    Tasks become routines. Simple status mails and the Monday report run on their own. Complaints and quotes over your limit wait for your OK, because that is how you set it up.

  3. 3

    Knowledge

    Learns from day one. The Company Brain grows: standard answers, pricing rules, who approves what. When Nadine goes on leave, the work does not stop.

  4. 4

    Growth

    When it gets more complex, you do not start over. Build a custom back-office app with its own interface on the Superkind platform, on top of what Superkind already knows. Built yourself or with the Superkind team, and it all belongs to you.

Send Superkind your first routine task

Start for free in Outlook, Teams or Slack.

A classic executive office chair, empty, with a single coral-orange piping line along one seam on the backrest

How the shift works

AI does not arrive as a single release that deletes a department. It changes work in four slow, overlapping moves. If you know the pattern, you can see where your own job sits in it.

  1. Task-level automation first - the AI takes over single, repeatable tasks (a classification, a draft, a lookup). The role stays, the mix of tasks changes. This is where most companies are today.
  2. Whole workflows next - several tasks that always went together get stitched into a routine: new mail arrives, AI reads it, pulls data from the CRM, drafts the answer, logs the call. A person approves and the workflow runs on its own. McKinsey calls workflow redesign the step that pays, not task-by-task automation.
  3. Role content changes - the role description is rewritten. The clerk is now responsible for exceptions, approvals and continuous improvement of the routine, not for data entry. Pay stays or rises; the PwC Jobs Barometer shows a 56 % wage premium for AI-skilled workers in 2024.
  4. New roles emerge - AI workflow owners, AI compliance leads, Company Brain curators, people who build and tune AI employees. The WEF projects 170 million new jobs by 2030, led by data, AI, cybersecurity and the care economy.

The pace is faster than earlier waves but not uniform. The IAB’s substitutability potential (the share of a job’s tasks a machine could technically do today) rose from 15 % of German employees in high-potential jobs in 2013 to 38 % in 2022. Potential, not reality: most employers still use a fraction of what is possible.

Which jobs are most exposed

The jobs that lose the most hours are the ones with the highest share of text-heavy, data-heavy, routine tasks and little face-to-face work. McKinsey finds that up to 54 % of expected job transitions in advanced economies sit in office and administrative work.

  • Office and administration - data entry, document processing, booking invoices, matching bank statements, scheduling, letter writing. Entire tasks move to AI; the person moves to approving and handling exceptions.
  • Simple customer service - order status, address changes, invoice copies, delivery questions. Up to 80 % of common issues could be resolved without human intervention, Gartner projects. Complaints and relationship work stay with people.
  • Translation and transcription - routine translations and meeting notes become near-free. Specialised, legal and literary translation remains with professionals.
  • Routine data analysis and reporting - the weekly report, standard dashboards and ad-hoc number pulls move to AI. Analysts shift to the harder question of why the numbers moved.
  • Sales admin and inside sales - lead qualification, CRM updates, quote drafting. The sales conversation stays with the human.
  • Parts of accounting and finance - invoice verification, bank reconciliation, dunning, month-end preparation. The decisions and the audit stay with people.
  • Parts of marketing and content - first drafts, social copy, newsletter variants. Strategy, brand voice and visual direction stay with people.

Even here, Goldman Sachs estimated only about a quarter to a half of the workload inside exposed jobs can be done by AI. The clerk still signs off; the sales rep still runs the meeting.

Which jobs stay and which emerge

The jobs least affected share one feature: they depend on being present, trusted and responsible in ways no model is.

  • Healthcare and care - nurses, caregivers, therapists, doctors. The International Labour Organization puts jobs in Germany facing full automation at only 5.5 %, and this cluster is at the far end of it.
  • Teaching, child care and social work - the Bundesagentur für Arbeit still lists shortages across these fields.
  • Skilled trades - electricians, plumbers, mechatronics technicians, carpenters. German Mittelstand companies list these at the top of their skilled-labour shortage.
  • Leadership and judgement roles - managing directors, team leads, department heads. The decision, the people call, the accountability do not transfer.
  • Specialised craft, design and creative direction - architects, designers, master craftspeople, strategy consultants. AI drafts; they decide.

New roles the WEF Future of Jobs 2025 report lists among the fastest growing: data analysts and scientists, AI and machine learning specialists, big-data specialists, information-security analysts, software developers, environmental engineers, education and care roles. Add to that the quieter new titles inside every company: the person who owns the AI employee for accounts receivable, the person who writes approval rules, the Company Brain curator. These roles rarely show up on a job board as such; they grow out of today’s department heads.

AI employees vs. chatbots vs. RPA

Three technologies get mixed up in the “AI replaces jobs” conversation. They do very different things to a role.

OptionWhat it doesEffect on rolesHosting / Data protection
SuperkindAI employees that read mail, reason over your systems and act in your tools (Outlook, Teams, Slack, SharePoint, DATEV, HubSpot). Routines run on their own; sensitive steps wait for your approval. 1,500+ integrations.Takes repetitive tasks off a role, so the person moves to exceptions, approvals and work with customers. No team ever “becomes” the AI.Hosted and processed in Europe, GDPR compliant, EU AI Act aligned (Regulation (EU) 2024/1689), no training on your data, DPA through the Trust Center.
Chatbot (classic)Answers in a chat window, usually scripted or trained on an FAQ. Does not act in your systems.Shaves a few routine questions off customer service; the rest of the role is unchanged.Varies by vendor; check the vendor’s pricing and data pages.
Copilot (chat alone)Drafts text next to a document or an email, mostly on request. The person still copies and sends.Faster drafting; roles unchanged.US-hosted by default; EU Data Boundary options at a surcharge, check with vendor.
RPA (robotic process automation)Clicks through fixed screen flows on behalf of a user. Breaks when the interface changes.Removes a few mechanical steps from a role; a developer owns the robot.Depends on where the server runs; check with vendor.
No AI (today)People do every step by hand, between email, Excel and the ERP.Role unchanged, slower, more exposed to the labour shortage.Only where your own systems run.

The practical answer to “will AI replace my job” depends mostly on what kind of AI your employer brings in. A chatbot adds buttons. An AI employee changes what you do all day.

Examples from daily work

Five concrete examples, two from Superkind customers, three from research. Superkind is the first.

  • Shared inbox at a Mittelstand service provider - the Inbox Clearer answers 70 per cent of routine mail from the service@ address on its own, pulling order status from HubSpot and invoice copies from SharePoint. The two colleagues who used to split that queue now handle complaints and new business.
  • Monday report at a Hanover software team - the Report Builder opens Excel, HubSpot and the DATEV export at 07:00 every Monday, writes the one-page summary in the SharePoint template and posts it in Teams. The management meeting starts with the report, not with building it.
  • Customer service at scale - Gartner projects that agentic AI will resolve 80 per cent of common customer-service issues without human intervention by 2029. Call-centre roles do not disappear; the mix shifts to tier-two and complaint work.
  • Invoice booking in accounting - a Mittelstand accountant we interviewed went from booking 180 invoices a day to checking 180 AI drafts and spending the rest of the day on month-end work she used to postpone.
  • Translation and subtitling - routine business translation moved to AI post-editing: translators do final-mile work and specialise in legal, medical or marketing copy where the error cost is high.

What it takes to make the shift work for you

Making the shift count, as a company or as a person, is not expensive. The cost sits in deciding and talking, not in software.

  • For the company - a free start at app.superkind.ai with up to EUR 150 in credits, no credit card. The Business plan starts at EUR 50 per month for the whole company with unlimited users (so you do not pay per seat); EUR 100 covers up to 1,000 tasks or 5,000 questions. Enterprise is a custom price with SSO, SCIM and audit log. Stand: October 2026. Compared to the fully loaded cost of a routine admin role (salary, employer contributions, tools, management time), the first results usually show up in the first weeks.
  • For the person - one hour to try it, two weeks to see where it helps your own work. Day one is still a browser and a free account. The PwC Barometer shows a 56 % wage premium for AI-skilled workers in 2024; it rewards the person who learned to use it, not the person who waited for permission.
  • For the team - clear rules: which steps need approval (sensitive mail, quotes over a limit, personnel decisions), which run on their own. One shared Company Brain so knowledge does not leave with people.

Data protection and EU hosting

The question “will AI replace jobs” carries a second question with it: where does the data behind the work live? Customer mails, invoices, personnel files, prices. Superkind keeps all of it in Europe. The AI models (OpenAI, Anthropic, Mistral, Google) are used from Europe, under GDPR, under the EU AI Act, with no training on your data. The works council and the DPO can read the DPA before any task is sent. For a lot of German companies, that is the difference between a pilot that spreads and a pilot that stalls. See our AI agents in the Mittelstand deep dive for the EU AI Act timeline and the EU AI Act compliance guide for the obligations.

How to start

Three steps for a department head who wants to know by the end of the quarter which parts of the team’s work AI should do, and which stay. Superkind is the shortest path.

  1. Pick one task on one desk, this week - the task everyone grumbles about on Monday: the shared inbox, the weekly report, routine quotes. Sign up for free at app.superkind.ai and give Superkind that one task in Outlook, Teams or Slack. First result the same day.
  2. Turn it into a routine, with approval where it matters - once the task runs cleanly, tell Superkind what should happen every week and where you want to approve. Thanks to the 1,500+ integrations, Outlook, Teams, SharePoint, HubSpot and DATEV are connected without new tools or custom code.
  3. Open the conversation with the team - say out loud which tasks you want AI to take, which stay, and what the person on the desk should learn and grow into. The research is clear: the people who grow in this period are the ones whose managers made the choice explicit. On the Superkind platform you later build your own apps on top of what Superkind already knows; it all belongs to you.

Conclusion

Will AI replace jobs? Mostly it replaces tasks, builds new jobs beside the old ones and leaves the people who stay with better work. Office, admin, simple service and parts of finance will feel it first; care, teaching, trades, leadership and skilled crafts least. The honest answer in Germany is not 300 million jobs vanishing or 5.5 per cent, it is both: a lot of tasks shift, a few whole jobs disappear, and the people who learn to run the AI next to them will be the ones your company counts on. The fastest way to see it on your own desk is to send Superkind one real task this week and compare the result with how you would have done it yourself. If you want to think further, read our AI agents in the Mittelstand piece and the AI agents and the labour shortage deep dive.

Frequently Asked Questions

Rarely a whole job, but a growing share of the tasks inside most jobs. The World Economic Forum projects 170 million new jobs and 92 million displaced by 2030, a net gain of 78 million. McKinsey calls the 57 % of US work hours that are technically automatable a potential, not a forecast. The share of tasks AI touches grows; what stays is the human judgement, the relationships and the responsibility.

Office and administrative work (data entry, document processing, basic case work), simple customer service, translation and transcription, routine data analysis and parts of sales administration. McKinsey finds up to 54 % of expected job transitions sit in office and admin. The jobs change more than they disappear: the clerk becomes the clerk who checks what AI drafted.

Jobs that depend on physical presence, trust, empathy and complex judgement: nursing and care, teaching and child care, skilled trades (electricians, plumbers, mechanics), creative and leadership roles, and specialised crafts. The International Labour Organization estimates only 5.5 % of jobs in Germany face full automation.

Faster than previous technology waves, but not overnight. The share of German employees in jobs with high substitutability potential rose from 15 % in 2013 to 38 % in 2022, according to the IAB (the research arm of the Bundesagentur für Arbeit). PwC finds AI-skilled workers already earned a 56 % wage premium in 2024. Most companies will see the change over three to five years, department by department.

Not on its own, and not the whole job. What leaves your desk first are the repetitive parts: typing data from an email into DATEV, pulling numbers for the weekly report, drafting the same six paragraphs again. What stays is the part your customers and colleagues ask you by name: judgement, trust, the decision. Superkind's view is 'complements, does not replace': the AI runs the routine, the person owns the result. Teams report the same thing in practice - the clerk still signs off; the sales rep still runs the meeting. The role changes shape, the headcount usually does not shrink, and the person who learned the AI is the one the manager asks for next.

The data behind every task (customers, invoices, personnel files) is why Europe-based AI matters. Superkind is hosted and processed in Europe, is GDPR compliant and aligned with the EU AI Act (Regulation (EU) 2024/1689), does not train models on your data and offers a data processing agreement through the Trust Center. The AI models (OpenAI, Anthropic, Mistral, Google) are used from Europe.

You can start for free at app.superkind.ai with up to EUR 150 in credits, no credit card. The Business plan starts at EUR 50 per month for the whole company with unlimited users, so you do not pay per seat. EUR 100 covers up to 1,000 tasks or 5,000 questions. Compared to the fully loaded cost of a routine role (salary, employer contributions, tools), the first results usually show in the first weeks. Stand: October 2026.

Related Articles

Henri Jung, Co-founder at Superkind
Henri Jung

Co-founder of Superkind, the AI employee teams write to in Outlook, Teams and Slack. Henri helps companies, from small teams to the Mittelstand, hand their routine work to AI - without new tools, consultants or an in-house AI team.

Give Superkind the task you grumbled about on Monday.

Start for free in Outlook, Teams or Slack. Or let us show you in 30 minutes how it runs in your systems.