Today
On site, the customer's contact quickly mentions what else she needs, and at the machine someone notices that something is wrong. It gets written on a scrap of paper, in a chat message to oneself or not at all. By the evening half of it is forgotten, and a promise to the customer stays open until they chase it.
An example
A site manager at Blank & Kamp speaks into her phone for 40 seconds on the way from the congress center to the hospital site: the congress center also wants glass cleaning in the foyer from December, quote by Friday, and the B 650 scrubber dryer at the hospital leaves streaks. Once she confirms, demi creates two tasks: the quote for inside sales, linked to the customer's framework contract, and a service order for the B 650 with its serial number. On Thursday demi sends a reminder about the quote.
Your data, wherever it lives
Voice notes sit on phones, in chats and on voicemail today; customers and orders live in the ERP and CRM, appointments in the calendar. demi reads, understands and links them, with your rights. The AI turns every note into a task with owner, due date and context in seconds, and you confirm it with one tap.
Where the data lives today
- Phone & voice
- Chats
- ERP
- CRM
- Calendars
How demi solves it
Three layers in this order: connect the data first, then the rights, then the AI.
- 1 · ConnectData Layer
This case needs data from phone and chat, the ERP, CRM and calendar. demi replaces nothing and links every voice note to customer, site, order or machine. The task lands in the task system you already use or in the owner's calendar.
- 2 · PermitPermission Layer
The note is yours until you confirm the task. After that only the owner and those who may see the customer or order anyway can see it. Anything that goes to the customer is always released by a person.
- 3 · ActIntelligence Layer
The agent transcribes the note, recognizes what needs doing, for whom and by when, and finds customer, order or machine in the ERP and CRM. If something is unclear, it asks instead of guessing. Before the due date it reminds the owner and tells you when something stalls.
How it runs
- 1
You speak the note into your phone, by app, chat or call.
- 2
The agent transcribes it and recognizes task, owner and due date.
- 3
It links the task to customer, site, order or machine.
- 4
You confirm with one tap or correct it by voice.
- 5
demi creates the task, reminds in time and flags it when it stalls.
What changes
Promises made between two appointments become tasks at once
Every task is linked to the right customer or order
Fewer scraps of paper and messages to yourself
What does it bring you?
Check it with your own numbers
Every starting value is an assumption you can change. The calculation sits right below, step by step.
Your numbers
Result
+192 hfreed per year
Assumptions and calculation
demi turns the voice note into a task with due date and context; you just confirm it.
8 min → 4.8 min per case
How we calculate
300 voice notes × 8 min × 40% × 1.0 (manual) × 12 months = 11,520 min = 192 h
Value per year, before demi costs
Freed per year+192 h
Assumption: labor cost of $48.62 per hour in US manufacturing in June 2026 (BLS), plus workplace, IT and overhead.
= Value per year, before demi costs
+$12,480
gained per year
192 h ÷ 1,760 h per role = 0.1 full-time roles
An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.
Combine with more cases and get the packageQuestions about this use case
How do I record a note?
However it suits you: in the app, as a voice message in the chat or by calling a fixed number. You do not need to learn commands; you speak as you would to a colleague. Speech recognition understands dialect and technical terms, and demi marks unclear passages.
What happens if demi cannot identify the customer clearly?
Then the agent asks, for example whether the congress center in Essen or the one in Bottrop is meant. It creates nothing until the link is clear. You answer by voice or with a tap.
Can I assign tasks to colleagues?
Yes. If you say who should do it, the task goes to that person with your note and the context. They see only what they need for the task and may see anyway. If they are away, demi suggests a stand-in.
Where is the recording processed and how long is it kept?
demi is hosted in the EU, and speech recognition also works with models in EU data centers; other regions only with your company's approval. You decide how long recordings are kept, for example only until the task is confirmed. After that the text and the task remain.
