One company, one AI gateway – AI in every department
An example company with 250 employees that develops its own products. We follow Anna from her first day to her last and see where the AI helps: at every workplace, at reception and in the contact centre, in HR, IT, purchasing, marketing and sales.
From the signed contract to the last working day
Fifteen stations in one continuous story. At each one: what the AI does, who decides, which systems are involved, and why it runs on the cloud AI or on the on-premise AI.
Show detailsHide details: From the signed contract to the last working day
- 1Join
The contract is signed
Cloud AI · names replacedThe AI does- Reads the contract and checks that nothing is missing
- Asks for missing details
- Starts every following step on its own
Who decidesHR confirms the start date
HR systemWhy cloud: Reading a contract precisely needs a highly capable model. Names and figures are replaced by pseudonyms first.
- 2Join
Accounts and access
Cloud AI · names replacedThe AI does- Creates the user account and the mailbox
- Adds the right groups and folders for the role
Who decidesIT approves unusual permissions
User accountsEmailWhy cloud: Routine work for the highly capable model; names are replaced before anything leaves.
- 3Join
Licences
Cloud AIThe AI does- Assigns the software licences the role needs
- Looks for unused licences first and re-uses them
- Only then proposes a purchase
Who decidesPurchasing approves new licences
Licence managementWhy cloud: Only roles and licence counts, nothing confidential.
- 4Join
Laptop, phone and number
Cloud AIThe AI does- Chooses the standard equipment for the role
- Places the order and tracks the delivery
- Sets up the phone extension
Who decidesThe manager approves the order
Ticket systemSupplier shopPBXWhy cloud: A routine order, nothing confidential.
- 5Join
The first day
Cloud AI · names replacedThe AI does- Writes the welcome email and the plan for the first week
- Books the introduction training
- Answers "how do I …?" questions at any time
Who decidesThe manager adds a personal note
TeamsIntranetWhy cloud: A warm, well-written welcome needs a strong writer.
- 6Work
Anna's everyday office work
Cloud AI · names replacedThe AI does- Drafts replies to her emails
- Turns notes into a finished presentation
- Builds Excel tables and evaluations from raw numbers
- Summarises meetings and lists the to-dos
- Finds the right document in seconds
Who decidesAnna checks, corrects and sends
EmailPowerPointExcelTeamsDocument storageWhy cloud: Good emails, slides and tables need the most capable model. Personal data is replaced by pseudonyms.
- 7Work
Marketing
Cloud AIThe AI does- Drafts newsletters, posts and website texts in the company's tone
- Translates them
- Evaluates which campaign worked
Who decidesMarketing approves before anything is published
WebsiteNewsletter toolStatisticsWhy cloud: Texts for the public: the strongest writer wins, and nothing is secret.
- 8Work
Sales and customer service
Cloud AI · names replacedThe AI does- Prepares quotes from the price list
- Updates the customer record after every call
- Reminds the team of open follow-ups
Who decidesSales approves every quote
CRMERPWhy cloud: Convincing quotes need a highly capable model; customer names are replaced.
- 9Work
Reception
Cloud AI · names replacedThe AI does- Answers the main phone number and puts callers through to the right person
- Registers visitors, tells the host and prepares the badge
- Books meeting rooms
- Sorts the incoming post and the general inbox
Who decidesReception handles everything unusual
PBXVisitor managementCalendarEmailWhy cloud: A natural conversation on the phone needs a highly capable model.
- 10Work
The contact centre
Cloud AI · names replacedThe AI does- Answers every call at once, day and night
- Solves standard requests on its own
- Suggests answers to the agent during a call
- Writes the call summary and updates the customer record
- Shows why customers are calling this week
Who decidesAgents take over anything complex or emotional
PBXCRMTicket systemKnowledge baseWhy cloud: Understanding callers and solving their request needs the most capable model.
- 11Work
Research and development
On-premise AIThe AI does- Searches lab notes, test results and earlier studies
- Compares test series and summarises the findings
- Drafts reports and patent texts
Who decidesThe researchers check every result
Lab notesTest dataPatent draftsWhy on-premise: Here the content itself is the secret. A formula cannot be hidden behind a pseudonym, so it never leaves the building.
- 12Work
IT keeps watch in the background
On-premise AIThe AI does- Reads the logs of all systems around the clock
- Tells normal from unusual
- Explains a problem in plain words and opens a ticket with the likely cause
Who decidesIT decides how to fix it
MonitoringServers & networkTicket systemWhy on-premise: System logs show how the network is built. They cannot be masked, so they stay inside.
- 13Work
Assessments and reviews
Cloud AI · names replacedThe AI does- Collects goals, results and feedback for a review
- Pre-fills supplier and security questionnaires
- Summarises the evidence for an audit
Who decidesPeople assess people – the AI only prepares
HR systemDocumentsWhy cloud: Names are replaced by pseudonyms; the judgement stays with people.
- 14Move
Anna changes department
Cloud AI · names replacedThe AI does- Adjusts permissions, licences and mailing lists to the new role
- Removes what is no longer needed
Who decidesThe new manager confirms
User accountsLicence managementWhy cloud: Routine work; names are replaced.
- 15Leave
Anna leaves the company
Cloud AI · names replacedThe AI does- Locks accounts and access on the last day
- Returns the licences
- Hands over the mailbox
- Requests laptop and phone back, releases the number
Who decidesHR and IT confirm
User accountsLicence managementTicket systemWhy cloud: Routine work; names are replaced.
What changes for Anna at her desk
No new application to learn. The AI sits in the tools she already uses and takes over the first draft of almost everything.
Show detailsHide details: What changes for Anna at her desk
The inbox
Twelve emails came in overnight. For each one a draft reply is waiting. Anna reads, corrects two, sends.
The meeting
The AI listens in. Afterwards the summary and the list of to-dos are already in the team channel.
The Excel table
"Show me sales by region for the last three years." The table and the chart are ready in a minute.
The presentation
Anna's bullet points become ten clean slides in the company design. She adjusts the wording.
The search
"Where is the framework contract with supplier X?" The AI finds it and quotes the paragraph.
The quote
A customer asks for prices. The AI prepares the quote from the price list; Anna approves it.
The strong AI for quality, the on-premise AI for know-how
Anna's company uses two AIs. A simple rule decides which one does the work: personal data can be replaced by pseudonyms, company secrets cannot.
Show detailsHide details: The strong AI for quality, the on-premise AI for know-how
Cloud AI The most capable models, such as Claude
- Why. The top cloud models are several times more capable than any model that can run on company-owned servers.
- Where that strength is needed. Emails and letters that must read well. Presentations and reports. Analyses in Excel. Long contracts and documents. Tasks with many steps. Conversations on the phone.
- How the data is protected. Names, addresses, account and tax numbers are replaced by pseudonyms before anything leaves the office. The AI works inside the EU.
- In Anna's company. 13 of the 15 stations: office work, HR, sales, marketing, reception, contact centre.
On-premise AI An open model on the company's servers
- Why. Nothing leaves the building. If needed, the server runs with no internet connection at all.
- Where that is needed. Wherever the content itself is the secret and no pseudonym can hide it: research results, formulas, product designs, patent drafts, system and security logs.
- What it can do. Less powerful than the cloud models, but strong enough to search, compare, summarise and draft.
- In Anna's company. 2 of the 15 stations: research and development, and IT monitoring.
Everyone works through the same AI gateway – it connects every system
Ten departments on top, the company's software at the bottom. In between: one AI gateway that checks every step.
- Email & calendar
- HR system
- CRM & ERP
- Ticket system
All 23 areas of a company where AI works today, from reception to the factory floor, and the software behind them: AI and existing software.
Your AI. Your data. In the EU.
Discuss a use case with us
20 minutes, without obligation: the situation is described, and we say what AI can realistically do, and what it cannot.
Open the form
- contact@ingnetworks.com
- Phone
- +49 1515 3851511
- Address
- Ingenious Networks GmbH
Eisenacher Straße 6
10777 Berlin, Germany