For agencies and companies with full inboxes
AI co-workers that ask before they decide.
On staffplane, agents answer mail, run processes step by step and put every important decision in front of a person – one click, even without an account. Every step is in the audit trail, and the platform runs on servers in Germany.
The problem
Why AI rarely makes it into daily work
The technology is there. What is missing is the trust to let it loose on real customers, applicants and partners.
The pilot never goes live
A chatbot answers test questions well. Still nobody releases it for real customer mail while it is unclear who decides and where that is recorded.
The inbox waits
Applications, enquiries, follow-ups: every mail needs someone to read it, sort it and start the next step – every single day.
Privacy slows everything down
Customers ask about data processing agreements, erasure and audit trails. Without clear answers, a demo never becomes an order.
How it works
The agent prepares. A person decides.
This is how a mail travels through staffplane – from the inbox to the decision.
- 01
The mail arrives
Every mail to a mailbox becomes a thread with a number, an owner and a state – in your team's shared inbox.
- 02
A process starts
A rule of the mailbox starts the matching process, for instance a review for every application. Where no rule matches, the agent chooses.
- 03
The agent prepares
It reads, summarises and asks back – within its role, its rules and its skills.
- 04
A person decides
Up to six answers to choose from, in the browser or by link without signing in. Then the process goes on.
Control
You keep the last word – and the evidence
Every safeguard is a setting, not a statement of intent. And whatever happens is in the audit trail.
Shadow mode
A new agent works silently beside your team: it writes its answers but sends nothing. You compare before you switch it on.
Decisions stay with people
A process waits for a person at every approval step. Their choice becomes part of the process.
Four eyes
If you want, a step needs two different people to agree – the first mail to a new contact included.
Kill switch
"Pause all agents" stops every agent of a Space at once, including running answers and tool calls.
Audit trail without contents
Who did what and when – through the interface, the API, MCP, GitOps and the channels. Message texts are not in it; the entries are chained and can be verified.
EU models and pseudonymisation
New Spaces use models hosted in the EU only. Names and contact details are replaced by placeholders before a model is called.
The offer
Two ways to your first AI co-worker
One offer for agencies that sell AI co-workers to their customers, one for companies that start on their own. Both begin with a demo on your case.
staffplane Partner – for agencies
Your own range of AI co-workers under your brand, without building or running a platform.
- 1A demo on a case of one of your customers
- 2Your partner Space: customer Spaces below it, your brand, your defaults
- 3The first customer starts in shadow mode
- 4staffplane takes care of operations, updates and evidence
Price on request. The platform measures usage per customer – the basis for your own prices.
First AI co-worker – for companies
An agent for one of your company's mailboxes: it sorts, prepares and asks back. Your team decides.
- 1A demo on a case from your mailbox
- 2Set up with a starter pack and the setup help
- 3Starts in shadow mode: it writes along, it sends nothing
- 4Runs with approvals, audit trail and kill switch
Price on request. If you work with an agency, it can run through them.
In development
A first use case: recruiting
A product of its own is being built on staffplane right now, for recruiters and companies with many applications. It uses the platform only through its public API and MCP – the way any agency could. It is not available yet.
This is how it is designed: applications arrive through a form, a mailbox or WhatsApp. The agent checks them against the customer's criteria, gives reasons and open questions, and puts the decision in front of a person – with four choices: invite, decline, put on hold or offer another position. The decision is made in the browser or by link without signing in; the agent proposes interview slots through Cal.com.
The platform's building blocks for it are built: decisions with choices, deciding by link, the first contact through WhatsApp templates, events to the product's own software, the call context for its own MCP servers and reading documents through the API. In this product a rejection is never sent by the AI: recruiting is a high-risk area under the EU AI Act.
Channels
Where your customers write
Every channel connection belongs to one agent. Voice messages are transcribed, pictures and documents come through. We do not provide phone numbers: WhatsApp runs on the company's own number.
- Available
E-mail mailboxes
Your own mailbox or an address the platform runs – with threads, owners and takeover by people.
- Available
Web chat
A widget for your website, embedded with one line. The AI notice appears before the first message.
- Available
WhatsApp
Through the official business platform on your own number. The agent opens a first contact only with an approved template.
- Available
Telegram
A bot of its own per agent, with voice messages, photos and PDFs.
- Available
Microsoft Teams
The agent answers in Teams; decisions appear as a card with up to six buttons.
- Available
Slack
A Slack app of its own per agent, answers as text.
- Planned
Phone
Answering calls is planned. Today the agent understands voice messages on the channels above.
Questions
Frequent questions before the start
Does the AI decide about people?
Only where you allow it. Every step of a process can be an approval by a person, by two if you want. What a person chooses is recorded in the process with name and time.
What if an agent writes nonsense?
A new agent starts in shadow mode and sends nothing. In operation every first mail to a new contact can wait for an approval, you check finished conversations in the post-hoc review, and the kill switch stops every agent of a Space at once.
Where is the data?
The platform runs on servers in Germany that we operate. New Spaces use models hosted in the EU only; before a model is called, personal details are replaced by placeholders. That is pseudonymisation, not anonymisation – detection does not catch every detail.
Do we need developers of our own?
No. Starter packs and the setup help get an agent going without programming. If you want more, use the public API, the platform MCP or Git.
What does staffplane cost?
Price on request. We discuss scope and terms once we know your case.
Can we offer staffplane under our own brand?
Yes: product name, logo, colour, support contact and the links to your legal texts are yours. A domain of your own for the interface is planned and not possible yet.
See it on your own case.
Tell us briefly what it is about. We show you the platform on an example close to your daily work.