Katasu Interview

You’re using AI, but aren’t seeing results.Does this sound familiar?

Katasu Interview uses AI to interview your teams and find where AI could improve their day-to-day work.

Start an interviewSee how it works

  1. Week 1Listen

    Everyone writes about their day-to-day work in detail.

    Work description

    When an inquiry comes in,
    check the details and

    Who checks the details, and how long does it take?

    route it to the right person.

    How do you choose who handles it?

  2. Week 2Map

    We create a report showing each step and how long it takes.

    Handling inquiries

    Per inquiry 10 min

    1. 1

      Admin

      Receive the inquiry

      2 min
    2. 2

      Admin

      Check the details and location

      5 min
    3. 3

      Admin

      Choose who will handle it and notify them

      3 min

    Assignment criteria　Request type and location
    If the person is unavailable　Ask the manager

  3. Week 3Delegate to AI

    See how much time you could save, then choose what to hand over to AI.

    Review and assign inquiries

    Time spent per inquiry

    10 min2 min

    Estimated saving: 8 min per inquiry

    AI
    Check the details and location, then assign the inquiry
    People
    Take incoming inquiries and review any that AI cannot assign
    Delegate to AI

What you’ll learn with Katasu Interview

See what can change,and by how much.

“AI will reduce your workload” does not tell you what will change or by how much.We look at how work gets done, then show you what could change and the estimated benefits.

  1. 01

    What can change?

    Tasks to delegate to AI, steps to eliminate, and decisions for people to make.

  2. 02

    What can you gain?

    How much time and money you could save, and how much more work your team could handle.

  3. 03

    What will it take?

    Setup and ongoing costs, the preparation needed, and what has to be in place.

  4. 04

    Where should you start?

    The first workflow to change and how to measure the impact.

How it works

Understand how your teams work.Find out what needs to change.

Open the sample report

  1. 01

    Listen

    AI interviews everyone, from managers to frontline staff, about the steps they follow, the exceptions they handle, and how they hand work over to others.

    Interview
    AI
    When a tenant reports a water leak, what do you do first?
    Staff member
    I check the contract, then give the tenant the maintenance company’s number. They’re the ones who arrange the repairs.
    AI
    Do tenants ever call you back after that?
    Staff member
    Quite often. The maintenance company tells them to ask us about the contract, so they call us again.

  2. 02

    Map

    We bring the answers together in one map showing who does what and how work passes between people. You can see where work waits, gets sent back, or goes through the same checks again.

    Workflow map

    Repair requests · based on interviews with 18 people

    From the first call to a confirmed repair date

    Loading the diagram…

    The tenant reports a leak to the property management company’s reception team → is told to call the maintenance company and makes another call → the maintenance company takes the repair request and asks the contracts team for details (waiting for a reply) → confirms the repair date → completes the repair.

    BottleneckWhat’s happening
    Referral to the maintenance companyThe reception team cannot arrange repairs, so the tenant has to make another call
    Requesting contract detailsRepairs cannot be scheduled until the contracts team replies

  3. 03

    Decide

    For each proposed change, we show the estimated time savings and costs, along with how we calculated them. You can decide whether to go ahead, hold off, or try one change first.

    Estimated impact

    Repair requests · 3 proposed changes

    Estimated impact

    Proposed changeBasis for estimateEstimated time savingsCost
    Use one point of contact for repair requests120 requests/month × 8 minutes spent on repeat calls16 hours/monthInternal process changes only
    Share contract details with the maintenance company120 requests/month × 12 minutes per inquiry24 hours/monthInformation-sharing system: JPY 30,000/month
    Eliminate duplicate data entry for address changes40 requests/month × 15 minutes of data entry10 hours/monthIntegration setup: JPY 200,000 upfront

    What we still need to confirm

    • Which contract details the maintenance company is allowed to see (subject to a review of the contracts)
    • Time saved does not automatically reduce staffing costs. We treat it as extra capacity to handle tenant requests.

Use cases

Practical changes start with listening to your teams.

We use interview findings to propose changes and estimate their benefits. The workflows and figures below are illustrative examples, not results from real companies.

  • Responding to emails

    You can’t reply to an email without asking someone who knows the details.

    What the interviews revealed
    The team spent an hour every morning reviewing emails and looking up information. Each reply took three rounds of internal back-and-forth. Two were just to find the right documents or check what they said.
    Proposed change and expected impact
    AI looks up the relevant documents and drafts a reply, then sends your team a consolidated list of questions that need human judgment. This could eliminate the two rounds of back-and-forth spent looking up documents.
    Internal back-and-forth per reply: 3 rounds → 1

  • Progress tracking

    You update the task board, but people still ask for updates.

    What the interviews revealed
    The team posted the same update on the task board and in chat every day. The people checking progress really wanted to know which tasks were stuck.
    Proposed change and expected impact
    AI uses work logs to update the task board and chat. This removes routine manual updates so people can focus on getting stalled work moving again.
    Routine manual updates per day: 2 → 0

  • Checks and approvals

    You repeat checks and approvals that have been in place for years.

    What the interviews revealed
    All 20 requests each day had to wait for a manager’s approval. Sixteen were routine requests covered by clear rules. Only four needed a case-by-case decision.
    Proposed change and expected impact
    Update the rules so the 16 routine requests can go ahead without approval. For the four exceptions, AI prepares the information the manager needs to make a decision.
    Requests needing manager approval per day: 20 → 4

FAQ

Frequently asked questions

How is this different from hiring a consulting firm?

AI interviews everyone, not just a few people. In about three weeks, you receive proposals based on those interviews, with estimated time savings and costs for each change.

How much time will our teams need to spend?

Each person answers the AI’s questions when it suits them. There’s no need to schedule meetings or prepare documents.

Which team can we start with?

You can start with the whole company, one team, or a single workflow. We’ll agree on the scope together in our first conversation.

How will you use the interview responses?

Responses are used only to understand the work and explore improvements. We’ll explain how we handle them when we speak with you.

Let’s start with a conversation.

We can help you decide which team to start with. You’re also welcome to simply ask how it works.

View a sample report