How you work
Two weeks in a team. Fourteen steps – or thirteen with Home. One state that counts afterwards.
The example on the home page explains the principle. This page shows the doing: who does what, with which tool, and what each step actually brings. It is the same case – critical customer feedback – only this time the way it runs day to day, with browser, AI, source and administration.
Owns the service knowledge. Manager in the workspace, leads the topic in Work.
Customer service. Contributor, works with a connected AI.
Customer service. Contributor, uses his AI without a connection at first.
Knows Git. Sets up the source and applies changes in it.
OrivelaIQ Home – Anna's knowledge source at OrivelaIQ. Created with one click, her AI writes into it – Tim is not needed.
With an own source: Tim sets it up and applies changes in it. 14 steps.
Once: the place comes into being
One afternoon, two people. After that, nobody touches the setup again.Half an hour, alone. After that, nobody touches the setup again.
The existing documents become the source
Tim creates a repository and brings in the documents that already exist: the customer feedback procedure, the escalation rules, two checklists. Markdown text, one SVG graphic. Nothing is rewritten.
What it bringsThe files stay yours. Nobody has to “migrate” into a new system.Anna creates her Home – one click
Anna creates her Home in OrivelaIQ. Her workspace “Service” comes into being with it, the first preparation runs by itself. Then she asks her AI: “Bring in our documents on customer feedback, escalation and the two checklists as articles.” The AI writes them into Home. Anna invites Mia and Jonas as contributors.
AIFour articles created in your Home, headings and structure taken from the originals. Nothing counts yet – you set the fixed state when you are satisfied.What it bringsNo repository, no tech colleague, no migration. Half an hour, and the place is there – with content.Create the workspace, connect the source, start the index
Anna creates the workspace “Service”, connects the source and starts the first index run. She invites Mia and Jonas as contributors. Both sign in with a one-time code by email.
What it bringsFrom now on there is one place where the service knowledge is readable and searchable – only for those who are meant to see it.Set the first fixed state
Anna reviews the working edition and sets it as the fixed state. She names the two reading states “Approved” and “In progress”.
What it bringsEveryone reads the same thing. Whatever changes in the source from now on appears only when Anna deliberately releases it.Connect your own AI
Mia connects her AI to OrivelaIQ once – a personal connection that belongs to her and that she can revoke at any time. The AI may do exactly what Mia may: read, and contribute to a topic once she is added to it.
What it bringsMia's chats know the state from now on. She no longer has to paste anything in.
Week 1: A question becomes visible
A real case, three people, two AIs – and nothing gets lost.
Classify a customer email
Ms Berger, a partner customer, complains. Mia asks her AI. The AI reads the fixed state through the connection and answers with reference to the article – and with a gap.
AIAccording to “Approved”: confirm receipt, reply within 48 hours. The article does not distinguish partner customers from others – whether something different applies to partners is not settled.What it bringsNo restart, no guessing. The AI says what counts – and names what is not settled, instead of inventing it.Feedback on the section
Jonas has the same case on his desk. He reads the article in the browser and leaves feedback on the section “Deadlines”: “Does this apply to partner customers too? Berger case.”
What it bringsThe question sits on the text, not in an email to three people. Everyone who reads the section sees it.Answer – and give the clarification a place
Anna sees the feedback in her workspace's feedback overview and answers: “Not settled. We'll clarify.” Then she creates the topic “Deadlines for partner customers” in Work, with a clarification point and a reference to the article in the fixed state. She adds Mia and Jonas to the topic.
What it bringsThe feedback stays open and traceable until it is really done. The clarification now has a place – separate from the state, but linked to it.A chat result becomes an entry
Mia has analysed the Berger case with her AI. She has the AI record the result as an insight in the topic, referring to the clarification point. The AI checks whether Mia may contribute to the topic, writes the entry and leaves a receipt.
AIInsight added to the topic “Deadlines for partner customers”: partner customers expect a personal reaction, not just a confirmation – two incidents in recent months. The state is unchanged.What it bringsThe result from the chat isn't lost – and it changes nothing about the state until someone decides.AI work arrives even without a connection
Jonas works with an AI that is not connected. He has it prepare his escalation experience as a portable work file and sends it to Anna. Anna imports it: plan, preview, apply – she sees exactly what goes into the topic beforehand.
What it bringsNo AI is excluded. And nothing goes in automatically – a person sees it first.The article shows what is in motion
Whoever opens the article now sees next to the text: “Open: deadlines for partner customers – one clarification point, two insights.” The topic's overflight shows which articles are touched. The article itself is unchanged.
What it bringsNobody works with a state any more without knowing that it is currently in question.
Week 2: Decide, change, release
Every step deliberate. The old state counts until the new one is released.
The decision – with reasoning
The team discusses it wherever it likes: in a meeting, in chat. Anna records the result as a decision in the topic: “Partner customers: the account owner reacts in person, within 24 hours.” The decision refers to the clarification point – which closes. The two insights hang on it.
What it bringsIn six months anyone can see why: the incidents, the experience, who decided. And: the old state still counts – deliberately.Change required – and applied
Anna sets the decision “change required” on the feedback. Tim changes the article in the source: a new section “Partner customers”, plus a reference to the decision document as its foundation. Anna records the change evidence on the feedback.
Honestly: with an own source this step needs someone who can operate it – Tim, or an AI with its own access there. OrivelaIQ never writes to a source silently.What it bringsThe source is changed deliberately, not on the side. And the path from feedback to change is complete.Anna tells her AI
Anna sets the decision “change required” on the feedback. Then one sentence to her AI: “Add the section Partner customers to the article Customer feedback, following the decision of day 8.” The AI writes into Home through its write tool – with Anna's separate approval – and names the decision as the foundation. The change evidence is thereby in place.
AISection “Partner customers” added to the working edition, foundation: decision of day 8. The fixed state is unchanged – it counts until you replace it.What it bringsFrom decision to source in one sentence. No Tim, no weekend, no Git. And still nothing counts until Anna releases.Review, release, close
Anna refreshes the index – a new working edition is created. She reads the changed article, sees the new foundation, and replaces the fixed state: “Approved October II”. On the feedback she confirms the verification. Jonas's feedback is closed.
What it bringsOnly now does the new state count – for everyone at once, including every connected AI, without anyone having to copy anything or send a notice.The same case, a different customer
Jonas has since connected his AI. New partner customer, new complaint. He asks – and gets the same answer Mia would get.
AIAccording to “Approved October II”: for partner customers the account owner reacts in person, within 24 hours. Decided by the team on day 8, based on two incidents and the escalation experience. Nothing is open on this.What it bringsTwo people, two AIs, one answer – with reasoning. That is the difference from day 1.
After two weeks
What has come into being – and what hasn't.
article changed – deliberately, with a foundation
feedback closed – with answer, decision, evidence
topic with decision, reasoning and a closed clarification point
fixed states – the old one stays traceable
lost chats, forgotten emails, silent changes
Where you find yourself
Four roles, four different shares of the work.
Reader
You read the fixed state, search, see what is open. You operate nothing. If you connect an AI, it reads the same as you.
Contributors like Mia and Jonas
You give feedback on the text and contribute to topics you have been added to – in the browser or through your AI. You never change the state.
Owners like Anna
You answer feedback, lead topics, make decisions and release the fixed state. This is the role that creates authority.
Tech like Tim
You set up an own source once and apply changes in it. With Home this role disappears – Anna and her AI take it over.
No tech role
With Home there is none. Anna creates Home, her AI writes, Anna releases. Anyone who prefers to stay in their own GitHub needs someone like Tim.
One branch that doesn't appear here: answers from outside. If feedback needs a specialist department or an external AI, it can go out as a package and the answer comes back traceably to the same spot – with visible authorship. The path afterwards is the same: decision, evidence, verification.
The next step
Your day 0 takes one hour.
You name a topic, we set up the state together, you connect your AI. Steps 1 to 4 we do with you – from step 5 on it is your everyday work.