VIENJA
Get a quoteFree 30-min call
Back to case studies
AI ProjectsAn AI product built by iPlus

Atrium — An AI Agent Office That Works Beside Your Staff

Atrium is an AI agent office: seven agents sort email and WhatsApp, draft quotes, enter orders and chase receivables, while people approve every step that matters.

Book a 30-minute callSee inside the product

On sample data · measured again before and after go-live

to answer a price request
12 min−87%
Before: 1 h 35 min
to enter a purchase order
3 min−83%
Before: 18 min
of manual work returned to staff each week
80 h
How it works

How Atrium works

The problem

  1. A customer asking for 40 servo drives by email passes through four people: front desk, sales for the quote, order entry, accounts for payment.
  2. Each step waits for someone to be free, so the customer waits over 1.5 hours for a quote.
  3. Every week, accounts spend 2.5 hours matching the bank feed to see who still owes money.
  4. Owners want AI to help, but fear it sending a wrong price out with nobody noticing.

What it does

  1. 01An agent covers each step: the Front Desk agent routes the message, the Sales Assistant drafts the quote from the price list and the customer’s past orders.
  2. 02Once the customer confirms, the Order Clerk records the order in kintone, and the Accounting Clerk matches the bank feed and drafts reminders.
  3. 03Quotes, large orders and reminders leave only after a manager clicks Approve.
  4. 04The office view shows which agent is doing what, and from which document.
  5. 05HR, weekly reports and translation work the same way: seven agents, each with a named supervisor.
Real product, sample data

Inside Atrium

01/10

The office floor: seven AI agents work beside staff, and every decision that matters goes to the approval desk

Rollout

Putting it to work

Rollout with iPlus

  • Weeks 1–2 · AssessmentPick the two or three highest-value workflows and measure today’s time and error rate.
  • Weeks 3–8 · BuildConnect Outlook, WhatsApp, kintone and QuickBooks; build the first three agents and run them in shadow mode on real data before low-risk work runs on its own.
  • Monthly · AI AdvisoryOptional, at least 3 months, to review the figures, tune the agents and add workflows.
  • You providePrice lists, letter templates, system access and a named owner for each step.

Running cost

  • SetupOne-time, fixed price, quoted after a 30-minute call.
  • AI APIUnder $150 a month for about 11,000 tasks (seven agents, 45 staff), billed to your own account with the provider.
  • HostingYour own cloud or a small server.

Where the AI gets its information

  • DocumentsPrice lists, contracts, policies and templates you add, each with an owner and a version.
  • Your softwareKintone, QuickBooks, Outlook and Excel Online, read live when needed.
  • No guessingWith no source the agent says so and hands over to a person, whose answer becomes new knowledge.
  • No trainingCustomer data is never used to train models.
Built withClaudeRAGZalo OAKiotViet APIMISA APIkintone APIGmail / Outlook APIGoogle SheetsNode.jsPostgreSQL

Frequently asked questions

Short answers to what companies usually ask before they start.

Ask anything else in 30 minutes
What happens when the AI is wrong?
Anything that must not be wrong, such as sending a quote, posting a large order or chasing a payment, waits for a person. Low-confidence items are held back.
Do staff need training?
Very little. Staff read a proposal and press Approve or Return; supervisors need one short briefing.
Does it work with our software?
Atrium connects to email, WhatsApp, kintone, QuickBooks and Excel Online. Other software is assessed in the AI Workflow Assessment.
Is our data safe?
Your data stays in your own systems and accounts; iPlus keeps none. It is never used to train models, and each role sees only what it needs.
What does it cost each month?
Beyond the fixed setup package, the API, paid directly to the provider, is estimated under $150 a month at about 11,000 tasks. The contract states the figure.
How long until it runs?
About 4–6 weeks for the first three agents, after a two-week AI Workflow Assessment that picks the work worth doing first.
What if we stop using it?
Your data and knowledge documents stay yours. Processes return to the old way of working, with no lock-in to a platform.
Can we start small?
Yes. Start with one or two agents, for example Front Desk and Order Clerk, and add more as the figures prove their worth.

Related projects

A faxed order is read and matched 3 of 3; the price is 12% over the old quote, so Orderly holds it for review.
AI Projects

Orderly — AI Agent for Purchase Order Entry

Orderly reads purchase orders from fax, email and PDF, finds the right part numbers and creates the order in kintone. Unusual orders wait for a person.

Luma answers a braces price from the 2026 price list and books Saturday in the chat, with no front-desk work.
AI Projects

Luma Care — Patient Care Chatbot for Dental Clinics

A chatbot that answers patients 24/7 on WhatsApp, LINE and the website, books on the real dentist calendar and hands clinical cases to a dentist. Every answer shows its source.

An employee asks about per diem: the answer is worked out step by step, with the exact clause beside it.
AI Projects

Marginalia — Policy Answers with the Clause Cited

Employees ask about leave, per diem or security rules in Slack or Teams and get an answer with the exact clause. When the rules are silent, it goes to HR, never a guess.