
STREIT · Hybrid Organisation · Customer service
How Tabea prepares orders at STREIT.
Tabea handles recurring preparation work in customer service: she reads order emails, matches customers and products, and creates order proposals. This gives the team more time for customers and special cases. MING Labs and STREIT developed the AI agent together.

Results reported by STREIT
More capacity in customer service.
Less manual order preparation creates room for customers and special cases. STREIT reports these results for that step in the workflow; the team remains in control.
- of orders prepared automatically
- Up to40–50%
- capacity gained per month
- Up to120 hrs
The next target: 80% automation rate. A future goal, not a result already achieved.
Reported by STREIT · September 2026. Customer-reported figures, not independently verified. The public source does not give the measurement period, order volume or capacity calculation. Time gained is not evidence of cost savings or additional revenue.
Read STREIT’s report ↗Public source checked on .
The task
Turning an email into an order proposal.
STREIT supplies office products, workplace technology and services. Its customer service team needs to understand incoming orders and match them to the right customers and products. Tabea was developed for this recurring work.
The division of work
Tabea prepares. People decide.
-
01 · Input
An order by email
Orders arrive at STREIT by email.
-
02 · Tabea
Match customers and products
Tabea reads and matches the order information.
-
03 · For the team
An order proposal
Staff receive a proposal to use in the next step.
Customer service at STREIT
Control and clarification
If a case is unclear or needs an additional decision, the team takes over. People remain in control.
Hybrid Organisation in practice
People and agents share a workflow.
Tabea has a defined task in customer service: preparing orders. The team uses her proposals and takes over when a case needs clarification or further decisions. For us, this division of work illustrates one part of a Hybrid Organisation: the task, the handover and human responsibility belong together.
MING Labs and STREIT developed Tabea together. Their collaboration combines MING Labs’ experience in building agents with STREIT’s knowledge of the processes and systems of mid-sized companies.
Explore Hybrid Organisation services →Apply the thinking to another workflow
Five questions for a first conversation.
These questions are a working aid from MING Labs. Start with one recurring task in your team.
Which task repeats?
Describe a specific piece of work and the result someone needs from it. A useful starting point is a frequent, clearly defined task where your team spends a lot of time on preparation.
What information comes in?
Identify the inputs, knowledge and systems involved. An anonymised example helps establish whether the necessary information is already complete or has to be gathered from different places.
What may an agent prepare?
Separate proposals, independent actions and required approvals. A draft that a designated person can review and approve before the next step is often a useful place to start.
Who takes over and owns the workflow?
Choose a typical exception. Identify who handles it and who owns the business process and technical operation. Also agree when the agent should stop and how it should hand over the case with enough context for that person to act.
How will you recognise an improvement?
Record today’s effort and the quality you need. Include review and rework. Later, compare similar tasks to see whether the time saved holds up after checking and correcting the output.
Copy the questions and add what you already know, with help from your AI assistant if useful. We can work through the open points in the conversation.
View the conversation template
Preparation for a conversation with MING Labs 1. Which task repeats? Describe a specific piece of work and the result someone needs from it. A useful starting point is a frequent, clearly defined task where your team spends a lot of time on preparation. [My notes or still open] 2. What information comes in? Identify the inputs, knowledge and systems involved. An anonymised example helps establish whether the necessary information is already complete or has to be gathered from different places. [My notes or still open] 3. What may an agent prepare? Separate proposals, independent actions and required approvals. A draft that a designated person can review and approve before the next step is often a useful place to start. [My notes or still open] 4. Who takes over and owns the workflow? Choose a typical exception. Identify who handles it and who owns the business process and technical operation. Also agree when the agent should stop and how it should hand over the case with enough context for that person to act. [My notes or still open] 5. How will you recognise an improvement? Record today’s effort and the quality you need. Include review and rework. Later, compare similar tasks to see whether the time saved holds up after checking and correcting the output. [My notes or still open] Already clarified: [Optional] My open decision: [Optional]
Questions about the STREIT case
Can I communicate with Tabea by email?
Tabea processes information from incoming order emails. STREIT’s public report does not say whether she has her own address for direct messages or sends email replies herself.
Who is responsible for Tabea?
The customer service team works with her order proposals and handles unclear cases or additional decisions. The public report does not identify who owns Tabea’s business role or manages her technical operation.
Does Tabea complete orders autonomously?
The report describes preparing order proposals for staff. People remain in control. It does not explain how each proposal is approved or processed further.
What do the reported hours mean?
STREIT describes time that becomes available for other work in the team. The calculation is not published; the figure is not a quantified cost saving.
Can other companies expect the same results?
The case shows one possible area of work. Whether an agent suits another workflow depends on its inputs, access, exceptions and review effort. STREIT’s figures are not a promise of results elsewhere.
The next step
Which task could give your team time back?
One recurring task from your team is enough to start. Together, we can explore what an agent could take on, who remains responsible and how you would recognise a useful result.
Discuss my workflow