The shift that works
while your office is empty.
And during the day, too. Zweite Schicht means second shift. We rebuild workflows in mid-sized companies with AI and run them, built into the systems you already have.
Book a callApproach and prices
- Start
- Workflow analysis, €4,900 fixed price
- For
- B2B, 50 to 2,000 employees
- Track record
- In Munich since 1998. AI in daily operation in energy technology, distribution and trade, among others.
What the AI gets done on a sample night 18:30 to 07:30
Environmental claims on ten domains in up to four languages checked against the EmpCo Directive, critical passages reworded, every passage logged.
Texts at scaleNew records from Salesforce matched: junk data filtered out before processing, unclear cases flagged for review.
Data in orderTexts for ten manufacturers built for the weekly rollout, eleven checks passed, published.
Texts at scaleRoutine checks run: interfaces, forms, availability. No deviations.
OperationsReport to the early shift: all done, three items need a human decision.
Handover
Signature
Day shift
And during the day, it works alongside your team.
The second shift does not stop at 07:30. During the day it takes the repetitive work off your team and prepares whatever a person has to decide.
Enquiry from Italy, received at eleven last night: read, classified, with a draft reply waiting for the southern sales team.
SalesNew lead from the web form created in Salesforce, company and industry added from the register, assigned to the right team.
SalesProduct adviser on the website: 61 technical questions answered in five languages, two of them handed to a person.
SalesReturn registered via the service portal, case created in Freshdesk, customer has the confirmation.
Tools
Both logs are examples, put together from real workflows at clients and in our own operations. The analysis shows how much your second shift can take on.
Findings
You have tried AI. Nothing has changed in day-to-day operations.
That is normal. A chat window answers questions, it does not do work. The work sits in your systems, in volume, with rules that have to be followed. Getting AI into that work takes hands-on engineering, not a prompt.
Marketing
Marketing maintains 4,000 products in four languages.
Every change to a data sheet passes through three pairs of hands. The translation waits at the agency, the approval with the product manager. By the time it is all online, the next data sheet has arrived.
Sales
Sales receives 60 enquiries a day.
Half of them are job applications, supplier pitches and questions from existing customers. The one promising enquiry from Switzerland is buried under 200 other emails on Monday morning.
CRM
The CRM holds 80,000 contacts, and nobody trusts them.
Duplicates, old addresses, fields filled in differently depending on who did it. Every campaign starts with a week in Excel.
HR
HR pays for 100 features and runs onboarding in Excel.
The HR system charges per user per month but does not know how your company works. So the most important part happens next to it: in a spreadsheet only one colleague understands.
Services
Four kinds of work the second shift takes on.
Three of them are in production at clients today; the fourth runs in our own operations and at some clients. Each has a fixed price, a defined scope and a figure you can measure the result by.
-
Rewrite, translate or generate entire text inventories: product texts from manufacturer data, environmental claims under the EmpCo Directive, eleven languages with your technical glossary. With volumes, turnaround times and a check that makes every passage traceable.
How it runs at a fuel cell manufacturerfrom €12,000fixed price per workflow, after the analysis -
Duplicates, junk data, empty mandatory fields, one manufacturer spelt three ways: we clean, structure and enrich, with rules you help set and a log for every record. The natural first step before any AI works on your data.
How it runs at a spare parts dealerfrom €12,000fixed price per dataset, after the analysis -
Read incoming enquiries, qualify them and route them to the right team. Prepare replies, create records in the CRM. A knowledge base that advises customers on your website in their own language and hands over to a person when it matters.
How it runs at a technology distributorfrom €12,000fixed price per workflow, after the analysis -
Software that adapts to your company, not the other way round. HR workflows, tasks and billing, service, quotes, portals, knowledge: built with AI, with AI inside, usable in your own terminology. Wherever standard software charges per user and still does not fit, or where none exists at all.
Where custom work winsFixed priceper tool, set in the analysis
Integration
Built into the systems you already have in place.
The work does not live in a new tool but between the ones you have. The second shift is the layer that connects them.
- CRM and salesSalesforce, HubSpot, Pipedrive, Microsoft Dynamics, Pardot
- ERP and inventoryNetSuite, SAP, Microsoft Dynamics 365, weclapp, Odoo, Sage
- HR and recruitingPersonio, SAP SuccessFactors, Workday, softgarden, d.vinci
- Product data and PIMAkeneo, Pimcore, manufacturer databases, Excel files
- CommunicationMicrosoft 365, Exchange, Google Workspace, Postmark, Brevo, Freshdesk
- Web and contentWordPress, TYPO3, Shopware, custom portals, Cloudflare
Your system is not on the list? If it has an interface, we can work with it. We have been doing that since 1998.
Custom work
Software that adapts to your company. Not the other way round.
Standard software charges per user per month and forces your workflows into its screens. Custom tools built with AI are now often cheaper than a few years of licences, fit exactly and actually get used, because they work the way your people already work.
Your workflow, not the software's. Who creates, who checks, who approves: the way it works at your company, not the way the vendor intended.
Built with AI, with AI inside. That is why it is ready in weeks and affordable. Reading documents, pre-filling fields, sorting what comes in, preparing decisions.
Used from day one. An interface in your company's own words, with no menus for features you do not need. No training day.
Yours, and always up to date. When the workflow changes, the tool changes, within days. No per-user subscription, no release cycle, no vendor doubling the price.
How it works
Six weeks from first look to live operation.
Not a consulting project that ends with a slide deck. A schedule with something running at the end of it.
- Week 1
Workflow analysis, three days with you
We watch where the work actually happens, who does it and which systems are involved. The result: a list of your workflows, ranked by effort today and savings tomorrow, with a fixed price per workflow. And a column headed "does not pay off".
- Weeks 2 to 5
Building the first workflow
We build in your systems, with your data. The new workflow first runs in parallel with the old one, with every output checked. Only when the figures are right does it take over on its own.
- Week 6
Handover and operation
Your team learns to work with the workflow and adjust it. From here the second shift runs: around the clock, on schedule, reporting to you. We keep it running.
Evidence
Three clients, three systems, three figures.
No market statistics, no invented quotes. Real clients with industry, size and systems, and workflows that run today.
-
10 domainsin up to four languages checked against EmpCo rules and reworded
Fuel cell manufacturer, listed
Around 400 employees, ten domains in up to four languages. Claims audit under EU Directive 2024/825, AI-assisted rewording with human review of every passage. Salesforce, SAP SuccessFactors and Freshdesk connected.
Read the case study -
3.3 million pagesfrom Salesforce data, manufacturer texts fully automated with quality checks
Spare parts dealer, around 50 employees
Independent dealer in industrial spare parts. Salesforce as the data source for 2,226 manufacturers in four languages, AI-researched texts in an automated weekly rollout, junk data filtered out, queries cut by 71 percent.
Read the case study -
+400% enquiriessince the relaunch, with AI advice, lead routing and NetSuite
Technology distributor, twelve locations
Around 2,000 employees, eight specialist markets across Europe. Knowledge base with a chatbot in five languages, enquiries routed by rules to regional sales teams, product data generated by AI from manufacturer documents.
Read the case study
All four cases, including our own operations, are under References: with figures before and after, the path in weeks and what went wrong.
Prices
Three stages, three prices, all published.
You start small and decide again after each stage. All prices net, plus VAT.
-
AI workflow analysis
€4,900fixed price, three days
- Review of your workflows on site or remotely
- List by effort and savings, with a fixed price per workflow
- An honest column: where AI does not pay off
- Data flow sketch for each workflow
Our commitment: a fixed price
Request an analysis -
Build per workflow
€12,000 to €40,000fixed price, binding after the analysis
- Built in your systems, with your data
- Parallel operation with every output checked
- Rollout and handover to your team
- The code belongs to you
Our commitment: a fixed price after the analysis
What determines the price -
Ongoing operation
from €2,900a month, cancellable monthly
- Monitoring of all workflows, reports to you
- Model updates and adjustment to new rules
- Response by the next working day
- Fixed allowance for extensions
Our commitment: cancellable monthly
What is included
Comparison
Not a consulting machine. Not a start-up.
In business since 1998, building with the AI tools of 2026. What that means by comparison, line by line.
| Criterion | Large consultancy | AI start-up | Zweite Schicht |
|---|---|---|---|
| First result | after months | after weeks | analysis in three days, first workflow in six weeks |
| Entry price | five figures, on request | subscription, often unclear | €4,900, fixed price, stated here |
| Who does the work | junior team, partner at the meetings | founders, until the next funding round | the same specialists, from first call to operation |
| Still around in five years | yes | unclear | since 1998 |
| Your data | corporate cloud | US start-up stack | you decide where it is stored |
| Who owns the code | licence, platform | the start-up | you, in full once paid |
Commitment
Data sovereignty is not a footnote. It is a sentence in every contract.
Before you sign, we show you which data goes where for each service.
In every contract with us, word for word
No customer data leaves your systems without your written decision, and none of your data ever trains a third-party model.
On your own servers. Open models on your hardware, when data must not leave the building.
EU data centre. Providers with data centres in Europe.
Cloud under contract. Cloud models under a data processing agreement, with no use of your data.
And the works council?
We build workflows, not surveillance: what a workflow logs is defined in advance, employee data is not used to steer people, and we supply the documents for a works agreement. How to involve the works council early is covered in the article Introducing AI with the works council.
Industries
Workflows from your industry, not off the shelf.
Four industries whose workflows we know: five workflows each, the typical systems and the objections raised there first.
Specifications, spare parts enquiries, item master data, technical documentation and service reports. With answers on export control, confidential designs and product liability.
Data sheets into the PIM, manufacturer price lists, bill-of-materials enquiries, enquiries by market and country, product discontinuations.
Environmental claims under EmpCo, service content in many languages, tickets from installers, warranty documents and partner data.
Time tracking without timesheets, time-and-materials billing, the shared inbox, quotes and tenders, client and project data.
Knowledge
Read up before you call.
Articles on the questions that come up in our first calls: costs, data, the works council, the EU AI Act and which workflow is worth it.
- 10 minutes
AI in business: chat window or workflow?
Why a chat window changes little in day-to-day operations, how to recognise a real AI workflow and five questions for checking your own use of AI.
- 10 minutes
AI project costs for mid-sized companies
Analysis, build, ongoing operation, model costs and internal effort: what an AI project costs a mid-sized company, with a worked example and eight questions to ask of any offer.
- 12 minutes
EU AI Act: what mid-sized companies must do
The EU AI Act is concerned not with the technology but with how it is used. Which typical workflows at mid-sized companies trigger which obligations, in one table.
All articles or the terms in the glossary
Questions
What purchasing, IT and your data protection officer want to know first.
Five questions that open almost every first call. All others are under Questions and answers.
What does it cost in total?
The analysis costs €4,900. The build per workflow costs €12,000 to €40,000, as a fixed price that is binding after the analysis. Ongoing operation costs from €2,900 a month. A typical first year with two workflows comes to between €60,000 and €100,000 net, plus VAT. Whether it pays off is set out in the analysis, with figures from your own operations.
Where is our data while you work?
In your systems. We build into them, not out of them. For the AI processing you choose per workflow between open models on your own servers, EU data centres and cloud models under a data processing agreement. The data flow for each service is on the data sovereignty page.
What happens if the AI makes a mistake?
Every workflow has rules, checks and a log. Whatever the rules cannot decide clearly, the workflow puts in front of a person on your side. Approval of texts, data and customer contact stays with you. That is not small print, it is the difference between a prompt and a workflow.
Who owns what you build?
You do. Code, rules, prompts and documentation pass fully into your ownership once paid. No platform, no licence, no subscription tying you to us. You book ongoing operation because it takes work off your hands, not because you have to.
Who do you not work for?
For companies with fewer than about 50 employees and no data or text inventories of their own, our analysis rarely pays off. We say so in the first call, not after the invoice. Nor do we build chatbots as decoration, or AI where an Excel formula will do.
Language
Working with us in English
This English version is for English-speaking decision-makers in companies based in Germany, Austria or Switzerland: an international management team, a parent company abroad, a works council that reads English.
We work in English or German, as you prefer. Invoices are issued in euros by WDM Webdesign München GmbH, a German company; prices are net, plus VAT. Our working hours are German time (CET/CEST). The sections on the works council apply to sites in Germany.
Enquiry
Tell us about the workflow that costs you the most time.
We will get back to you on the next working day with a first assessment and a suggested time for a 30-minute call.
Prefer to book a fixed slot straight away? Book a 30-minute call directly