Skip to content
← Back to guide Getting Started With AutoServa

GuluStock and AutoServa · GuluStock and AutoServa: User Guide

Getting Started With AutoServa

VYROX Wiki · Updated 23 September 2026 · https://www.vyrox.com/wiki/autoserva-getting-started

Getting started with AutoServa means checking that one of your workflows is a good fit, contacting VYROX, and moving through a staged engagement that begins with a decision audit and ends with a supervised AI operator running your queue. This guide is for business owners and managers in Malaysia and elsewhere who are considering AutoServa.

Step 1: Check if Your Workflow Is a Good Fit

Any three of these signs usually justify a decision audit:

  • The same kind of case arrives most weeks of the year.
  • A second person already reviews the output.
  • Two competent people sometimes decide the same case differently.
  • The queue slows when one particular person is on leave.
  • Most of the work is reading documents and applying a rule to them.
  • You can point to the last hundred cases and their decisions.

Important

Any one of these means it is better to wait: volume is low enough for a checklist; nobody can say how the decision is made; a system migration or new regulation is about to change the workflow; the real problem is bad upstream data; the work has no judgement in it; or the owning team has not agreed they want this.

Step 2: Contact VYROX

  1. Send a WhatsApp message to +60 19 688 3338 (contact is by WhatsApp only).
  2. Describe the workflow, its rough volume and the systems it touches.
  3. Agree a working session to scope the decision audit.

There is no email enquiry channel and no contact form. Replies usually come the same working day, and you can ask questions before you book anything.

Step 3: The Delivery Stages

StageTimingWhat Happens
ObserveWeek 1Watch a real queue with the people who work it
CaptureWeeks 1 to 2Write down the unwritten rules, including exceptions
CodifyWeeks 2 to 3Produce an explicit decision framework
Fine-TuneWeeks 3 to 5Train the operator on your material and vocabulary
RunWeeks 5 to 8Shadow running, then supervised live running with humans releasing
ProveFrom week 6Every decision carries its evidence
ImproveContinuousCorrections feed back into the framework

Nothing is automated until the operator can already reach your team's answer on its own.

Roles Your Side Provides

RoleTimePurpose
Process Owner2 to 3 hours a week, weeks 1 to 3Settles disagreements and signs off the framework
The Person Who Decides3 to 4 hours a week, weeks 1 to 3Explains how cases are really decided
IT or Systems LeadA few hours, weeks 2 to 4Sets up the operator's account and access
Compliance or Risk LeadOne session plus a week 5 reviewConfirms guardrails match policy

What to Prepare

  • Real documents the work produces, including awkward cases.
  • Past cases with the decisions your team made.
  • Written standards: templates, terms, tolerances and service levels.
  • The escalation rules: who signs off what, at which threshold.
  • Scoped access to the systems in play: read-only first, with write access only from the supervised live run.
  • A few hours a week from the person who actually decides, for about three weeks.

You do not need a clean labelled dataset. If your process is not documented anywhere, that is the normal starting point: writing it down is the first thing the decision audit produces.

Pricing Stages

AutoServa is priced per workflow, not per seat, and there is no published price list.

  • Decision Audit: fixed fee, two to three weeks. You keep the written framework whether or not you continue.
  • Design and Deployment: quoted per workflow, five to eight weeks.
  • Ongoing: optional monthly support with quarterly framework review, model migration and a named engineer.

What Drives the Cost

In rough order of impact:

  1. How much of the decision framework already exists.
  2. How many distinct case types the queue contains.
  3. How many systems the operator must touch.
  4. How strict the governance requirements are.
  5. Where it has to run (cloud or local).
  6. Case volume once live.

Four Commitments

  • No seat licences.
  • No lock-in on your own material.
  • No charge to move models.
  • No markup on hardware.

Measuring Whether It Worked

The baseline is agreed during the decision audit, usually a mix of throughput, how often a case needed rework, and how often two people would have decided it differently. Results are compared with your own prior numbers, not an industry benchmark.

After the first workflow is live, nothing is added automatically. Further workflows are added only as the deployed ones earn trust, and each is faster to tune because the vocabulary, standards and precedent are already in place. Stopping at one queue is a legitimate outcome.

Frequently Asked Questions

How Long Until the Operator Is Live?

A supervised live run is typically reached in five to eight weeks.

Do I Have to Continue After the Decision Audit?

No. Each stage is committed separately, and you keep the decision framework from the audit.

What Happens if We Stop?

You keep the framework, precedent library and evidence trail in an open format, and VYROX access is removed.

Who From Our Side Matters Most?

The process owner and the person whose judgement the team actually defers to, often not the manager. They need roughly two to four hours a week each during the first three weeks.

Related Guides

  • AutoServa Overview
  • AutoServa delivery process
  • AutoServa pricing
  • AutoServa FAQ