Whitepaper for decision-makers

Agentic AI im Mittelstand. What AI agents do, where they pay off and how to get started safely.

Agentic AI in mid-sized companies, explained for management, finance and IT leaders: how to put AI agents to work and what to watch out for.

14 pages
Three tools for your organization
Free PDF, in German
Get the whitepaper

Download it right away. We'll also email you the link. The PDF is in German.

4.9 out of 5OMR Reviews

What's inside

Fourteen pages of know-how. Tools and rules you can use right away.

One page per idea, explained with examples from mid-sized companies. All chapters are listed below.

Ausschnitt von Seite 4: die Autonomie-Leiter mit den ersten Stufen Antwort zuliefern, Arbeitsschritt vorbereiten und Entwurf im System
Tool · Page 4

Autonomy ladder

Six levels, from supplying an answer to changes in your core system within fixed limits.

Ausschnitt von Seite 6: die Scorecard mit den Merkmalen Aufwand heute, feste Regeln, Fehlerfolgen und Wissen
Tool · Page 6

Scorecard for prioritizing use cases

Six criteria and a points system show which use cases are a good place to start.

Ausschnitt von Seite 8: die Business-Case-Vorlage mit 8.730 Euro Ergebnis im ersten Jahr und der Tabelle der Posten
Tool · Page 8

Business case template

The math for your first AI agent, with every assumption, worked through an example.

Ausschnitt von Seite 10: die fünf Leitplanken, beginnend mit Rechten, Freigaben und dem Schutz vor eingeschleusten Anweisungen
Security · Page 10

Five security guardrails

Permissions, approvals, safeguards, budgets and logs. Plus the essentials on liability, the EU AI Act and data protection.

All chapters

  1. 2 When AI doesn't just answer, but acts
  2. 3 From AI chat to AI agent
  3. 4 How independently an agent may work
  4. 5 Where AI agents pay off: eight areas of use
  5. 6 Finding the right use case
  6. 7 What AI agents deliver
  7. 8 The business case for your first agent
  8. 9 The most common concerns and what helps
  9. 10 Using them safely: guardrails for AI agents
  10. 11 What companies need to get started
  11. 12 Ten recommendations and a 90-day roadmap
  12. 13 How SCALAN supports you
  13. 14 Sources and glossary
Tobias Veit-Seidl

Tobias Veit-Seidl, founder of SCALAN
The whitepaper draws on our customer projects, four interviews with executives at mid-sized companies from August to October 2026, and sources such as Bitkom, Destatis, McKinsey and Gartner.

Get the whitepaper

Request it for free. And start reading right away.

Once you submit, you can download the PDF right away. We'll also email you the link, so you can share it with management, finance and IT. The whitepaper is in German.

PDF in German, 14 pages, as of November 2026
Autonomy ladder, scorecard and business case template
With sources from Bitkom, Destatis, McKinsey and Gartner

A result up front

Even your first AI agent pays off. In the example, the rollout pays for itself after 2.7 months.

A shared inbox receives 50 customer inquiries every day. Reading and answering them costs around €23,100 a year today. With an AI agent, your team only checks its drafts and sends them.

The full calculation, with template, is on page 8 →
over 13,500
invoices from which an AI agent builds the inventory list. Built in a few days.
In practice: a manufacturing company with high protection requirements, customer project on SCALAN
Model calculation · Shared inbox
Today€23,100 a year
With AI agent, year 138% less
38%
lower costs in the first year
€8,730 after all costs
49%
less from the second year on
€11,230 a year
2.7 months
until the rollout has paid for itself
from the monthly surplus

Model calculation from the whitepaper, page 8: 50 inquiries a day on 220 working days, 3 minutes per inquiry today, 1 minute to check and send with an AI agent, €42 per hour. Platform, AI usage, rollout and support are deducted. The figures are guide values; we work out the actual costs for your company in a call.

Who it's for

Written for decision-makers. Straight to the point.

The whitepaper follows the questions that management, finance and IT leaders ask before their first AI agent.

Management · Pages 2 to 5, 9 and 12

Where AI agents pay off and how to start safely

What sets AI agents apart from AI chats, eight areas of use, the most common concerns with answers, and a path to live operation in 90 days.

Finance · Pages 6 to 8

What your first AI agent costs and saves

The scorecard for choosing a use case, six everyday benefits and the business case template you fill in with your own numbers.

IT · Pages 4, 10 and 11

What permissions an AI agent gets and who controls it

The autonomy ladder, five guardrails, the essentials on the EU AI Act, data protection and the works council, and what to prepare before you start.

It's not a technical guide for developers. You'll find that on our tech blog →

FAQ

Questions about the whitepaper.

Anything else? Email us at hello@scalan.ai.

What does the whitepaper cost?

Nothing. You enter your name, email address and company and can download the PDF right away. We also email you the link. You only get further emails if you tick the box. The whitepaper is in German.

Who is the whitepaper written for?

For management, finance and IT leaders at mid-sized companies who want to get started with AI agents. It explains the basics in everyday language, without technical detail, and takes about 20 minutes to read.

What's the difference between an AI chat and an AI agent?

An AI chat answers questions and drafts texts; the actual work stays with people. An AI agent handles a recurring task from start to finish according to your company's rules, for example reading incoming emails, recognizing the request, looking up information and drafting a reply. A person checks the result before it goes out.

Does an AI agent pay off in the first year?

In the whitepaper's example, yes. An AI agent for 50 customer inquiries a day cuts the cost of the task by 38% in the first year, around €8,730 after all costs. The rollout pays for itself after about 2.7 months. What matters is how much working time the task takes up today. That's why the whitepaper recommends measuring volume and duration for two weeks first.

What is the autonomy ladder?

The autonomy ladder describes in six levels how independently an AI agent may work: from supplying an answer, through drafts that a person checks and sends, to changes in core systems within fixed limits. To get started, the whitepaper recommends levels 1 to 3, where a person checks every result.

What happens to my details?

Wir nutzen sie, um dir das Whitepaper zu schicken. Weitere E-Mails bekommst du nur mit deiner Einwilligung, abmelden kannst du dich jederzeit. Einzelheiten stehen in der Datenschutzerklärung. [Prüfen: Text mit dem Datenschutz abstimmen, Speicherort in HubSpot ergänzen]

Who is behind the whitepaper?

It's published by SCALAN, a platform for agentic process automation by MTC Solutions GmbH, based in Munich. It draws on customer projects, four interviews with executives at mid-sized companies and public sources such as Bitkom, Destatis, McKinsey and Gartner.

What's next

Three ways forward after reading. All free, you decide how fast.

Which one fits depends on how concrete your task already is.

In three minutes

Agentic AI Check

15 questions on use case, accountability and rules. You see your result right away. The check is in German.

Start the check →
Live · 10 participants max
Live every Tuesday

AI Office Hours

45 minutes in a small group to ask questions and get your bearings. Free, no prior knowledge needed, in German.

Find a session →
Personal call
20 min
✓
Look at your task together
✓
Find the right approach
✓
Agree on next steps
20 minutes

Personal call

We look at your task and show you the right approach: a ready-made AI agent, a custom automation or consulting.

Book a call →