How we work · Inneraktive

Five steps.
Each one clear.

What we do at each step, what you get from it, and what it costs. Shown on one real process, from the first call to the first monthly report. Pick your industry on the screen.

  1. 01Walk-throughFree · 30 min
  2. 02Discovery1–4 weeks · fixed fee
  3. 03ConnectFixed scope
  4. 04BuildFixed price · phased
  5. 05RunMonthly
01 Walk-through

A straight answer in thirty minutes.

What we do
  • You walk us through one process, step by step.
  • We give our read: what AI can fix, and what to leave alone.
  • No sample data and no system access.
What you get
  • One of three answers: a Sprint proposal, not now, or not a fit
  • If it is a yes, a fixed-fee proposal within two days
Time · costFree · 30 minutes, on a video call
02 Discovery

A map and a fixed price, before anything is built.

What we do
  • We sit with the people who do the work.
  • We map the process as it is today, and with AI.
  • We mark every step a person must approve.
What you get
  • Requirements, in your language
  • A process map, today and with AI
  • A firm Build price, fixed and phased
  • A Run proposal with the monthly fee
Time · cost1 to 4 weeks · fixed fee, credited to your Build within 120 days
03 Connect

Every system reads the same record.

What we do
  • We find the record every system refers to. Here, the matter.
  • We join your systems around it.
  • We replace nothing. Your people learn nothing new.
What you get
  • One matter number in every system
  • Links that hold, ready for the work to run on its own
Time · costFixed scope
04 Build

Forty minutes of typing becomes four.

What we do
  • We agree the data model, integrations and approval points.
  • We build inside your systems and test on your own records.
  • We launch with your people.
What you get
  • A live workflow inside your tools
  • A person approving every step that matters
  • The Run toolkit: logs, a review queue, a kill-switch and seven more
Time · costFixed price, phased · never more than half at launch
05 Run

It keeps improving, and it's yours.

What we do
  • We watch it every day.
  • We improve it every month.
  • We report against the targets we agreed.
What you get
  • A monthly report
  • Improvements, shipped
  • Ownership of the code, the data and the logs
Time · costMonthly
New matter intake · Matter 2213
Today
40min per matterAs you described it
StepSystem · recordTimeStatus
Discovery

Sprint or Full,
sized to your work.

One to four weeks at a fixed fee. Which one depends on how much of the business the work touches.

SprintFull
ProcessesOne processSeveral processes
SystemsOne or twoThree or more
RecordsStandard business recordsRegulated records, such as client, patient or financial data
Whatever you decide afterwards, these four are yours
01Requirements

In your language, with the numbers the system will be measured on.

02Process map

Today and with AI, with every human approval and the critical step marked.

03Firm Build price

Fixed and phased, with assumptions and exclusions stated.

04Run proposal

What we monitor, improve and answer for, and the monthly fee.

The readout takes 60 to 90 minutes, and the memo is yours to keep.

Fee credited against your Build if you go ahead within 120 days.

Discovery readout · Sample

See a Discovery readout before you book one.

Download the sample readout
Example
Online retailer
ProcessOrder problems and customer updates
Today3 systems · 4 people · about 9 hours a week
With AI1 approval per refund · about 1 hour a week
Build priceFixed, phased · stated in the readout
Run proposalMonthly · what we watch and report on
The Run toolkit

Every Build ships
with ten ways to stay in control.

All of it lives in your accounts and your repository, so you could run it without us if you ever had to.

Order problemsRunningDraft onlyOff · working from the runbook
05
01Execution log
  • Address problem · Order 4471SentDraft queuedPaused
  • Late courier · Order 4480SentDraft queuedPaused
  • Refund request · Order 4466ReviewDraft queuedPaused
03Review queue

250waiting with the team leaddrafts waitingnothing running

04ApproveEditReject
02Monitoring

All clearAll clearPaused by you

Alerts go to #ops-alerts

08Cost meter

27% of this month’s AI budget

07Evaluation

Quality 94 / 100 this week

06Audit log

10:04 Approved by the team lead · Order 4471

Export
09Report

October report, built from the logs

Open
10Runbook

How to run itHow to run itFallback: the steps by hand

Read
It all runs in your own accounts
  1. 01
    Execution log

    Every run recorded: inputs, outputs, decisions and cost.

  2. 02
    Monitoring and alerts

    Failures, delays and odd volumes flagged to a channel you choose.

  3. 03
    Review queue

    Anything the system is unsure of goes to a named person.

  4. 04
    Approval step

    Approve, edit or reject, in your own tools where possible.

  5. 05
    Modes and kill-switch

    Auto, draft-only or off. One click, no engineer needed.

  6. 06
    Audit log

    What the AI did and who approved it, per record. Exportable.

  7. 07
    Evaluation harness

    Output quality scored every week and before every release.

  8. 08
    Cost meter

    AI usage per workflow, from your own accounts, with a budget alert.

  9. 09
    Report generator

    The one-page monthly report, built from the logs.

  10. 10
    Runbooks

    How to run it, and how to fall back to the old way.

Governance

Built for the rules
you work under.

  • Policy
  • Audit logs
  • Data residency
  • DPA set-up
  • Staff guidance

Included in the baseline Build for regulated clients.

What we hold ourselves to
  1. 01Never a Build without a paid Discovery.
  2. 02Never a Build without a Run line.
  3. 03Never two architectures, where one is the option we advised against.
  4. 04Never a website or app with no operations component.
  5. 05Never a Run contract without the ownership clause.
Questions

Questions
about the steps.

Ask anything else in your Walk-through →
01How long until something is live?

It depends on the process and the systems involved. Discovery ends with a firm Build plan and dates, so you know before you commit.

02What if we only want Discovery?

That is fine. Discovery is a fixed-fee step, and the four outputs are yours to keep whatever you decide.

03Do we have to change our systems?

No. We join the systems you already run around your key record. Nothing is replaced, and your people learn nothing new.

04Who owns what you build?

You do: the code, the data, the logs and the runbooks. The ownership clause is in every Run contract.

05Who is on the Walk-through call?

Avin, our founder, with the person who owns the process on your side and, if possible, one person who does it every day.

06What do you need from our IT or vendor?

An hour with your IT lead and, for systems a vendor runs, their partner. We need an integration user and API or export access to each system. All of it is agreed in Discovery, before anything is built.

Start here

Step one is
thirty minutes.

Show us one process. We'll tell you whether AI can fix it, and if it can't, we'll say so.

Or try the free five-minute interview: can AI run your process? →
Prototype
Thanks. The calendar opens on the next step.

In the live site this hands over to the booking calendar with your details filled in.