WareBot, Chat-Driven Warehouse AI for AutoCount

Run your warehouse by chat. No menus, no training.

WareBot is a chat-driven warehouse assistant for AutoCount. Your team types, speaks or scans, and WareBot checks stock and runs eight warehouse operations against your live data, every change confirmed with one tap before it commits. The four screens your pickers used to navigate become one sentence.

✓Runs on live AutoCount data ✓Text · voice · barcode ✓Zero staff training

Who This Is For

Who on the warehouse floor actually benefits from WareBot?

WareBot replaces screen navigation with a sentence. That helps every role in the warehouse, but it changes the day completely for the four who touch stock most.

Picker / Packer

Voice or text queries while their hands are full. Hands-free stock checks at the rack, then a barcode scan to confirm the right item, all without putting the box down.

Warehouse Supervisor

Quick status checks and transfers without opening a dashboard. Ask where stock sits, move it between locations, and keep walking the floor.

Stock Controller

Fast stock-take initiation for cycle counting and clean transfer management, started by a sentence instead of a multi-step form.

New Staff

Zero learning curve on common operations. No menu paths to memorise, no command syntax, just plain language in English, Malay or Mandarin from day one.

What It Replaces

Every line on the left is a habit your team can drop this month.

The old way is not broken software, it is wasted motion: screens, forms, app-switching and miskeyed codes. WareBot collapses each of them into one instruction.

Before

The Old Way

✕Navigate four screens just to check a stock balance.
✕Open the transfer module and fill in a form to move stock.
✕Train every new hire on menu navigation before they are useful.
✕Switch apps for every different task.
✕Miskey location codes and chase down the resulting errors.

After

With WareBot

✓“How much A100 do we have?” — answered instantly.
✓“Move 50 units of A100 from WH-A to WH-B” — one sentence, one confirm.
✓Natural language from day one — nothing to train.
✓A single chat interface for every warehouse task.
✓Scan the barcode mid-conversation — no codes to mistype.

Capabilities

How does WareBot understand what a warehouse worker means?

Five capabilities work together so a worker can ask the way they would ask a colleague, and get an answer that streams back in real time.

Natural-Language Interaction

Type or speak commands in plain language. WareBot understands warehouse terminology without training and works in English, Malay, Mandarin and more. Every message is tagged by input mode, text, voice or scan, so conversation history keeps its context.

Barcode Scanning

Scan an item barcode alongside any message to give WareBot instant context. It resolves barcodes and sub-codes to the full item, code, UOM and description, and records the scan type in the conversation history.

Cloud Speech-to-Text

Record a voice note and WareBot transcribes it server-side through a Whisper-compatible service. It accepts m4a, mp3, mp4, mpga, wav, webm and ogg, and returns the transcript with a detected language and confidence score.

Real-Time Streaming

Responses stream progressively, so there is no waiting for a full answer. Visual progress indicators show each tool running, and confirmation and audit events stream inline. A loop guard caps each turn at five tool-call rounds so the AI never runs away.

Multi-Turn Memory

Session context is held for 30 minutes on a sliding window. WareBot remembers earlier queries, so a follow-up like “What about WH-B?” just works. History includes tool calls and results, preserving the full reasoning context.

Flexible AI Backend

Run on a cloud model like OpenAI GPT-4o, or self-host an OpenAI-compatible model such as Qwen or LLaMA. Base URL, model name, API key and the Whisper model are all configurable, so inference can stay on your own infrastructure.

Warehouse Operations

Which warehouse operations can WareBot run for you?

Stock and inventory queries execute on the spot with no confirmation. Every write operation runs against your live AutoCount data, pauses for a one-tap confirm, and returns a document number when it commits.

Check stock balance
By item code, optionally by location. All UOMs returned. No confirmation.

Look up an item
From any barcode or sub-code, code, UOM, description. No confirmation.

List locations
Every available warehouse location, on demand. No confirmation.

Write Operation	What It Does
Create stock transfer	Move stock between locations.
Create stock request	Inter-location requisition, with optional notes.
Create stock take	Record a physical count for cycle counting.
Create stock issue	Issue stock to a person or department.
Create stock receive	Receive stock into a location, with an optional PO number.
Update customer pick	Mark an invoice's WMS status as Picked.
Update customer pack	Mark an invoice's WMS status as Completed.
Create credit note	Multi-line credit, with a reason.

How It Works

Nothing changes your stock until a person taps confirm.

Speed on the read, safety on the write. WareBot answers questions instantly, but every operation that changes data passes through a confirmation gate built for a busy floor.

1

Ask

Type, speak or scan, in any supported language.
2

Review

WareBot shows the full operation: what, how many, from where, to where.
3

Confirm

One tap to confirm or cancel. Tokens expire in five minutes.
4

Commit

The token is consumed once, no replay. The change writes to AutoCount.
5

Document

A document number (DocNo) returns via an audit event.

No accidental changes. A confirmation token lives for five minutes and is consumed the instant it is used, so a stray tap, a re-sent message or a stale screen can never commit the same operation twice. The audit event with the document number is your proof of every change.

Security & Access

WareBot can only do what the person using it is already allowed to do.

There is no new permission layer to manage. WareBot inherits your existing Bosun WMS authentication and your existing AutoCount user permissions, enforced at the moment of execution.

✓

Same Token Auth as Every Endpoint

WareBot uses the same token authentication as all Bosun WMS endpoints. Invalid or expired tokens are rejected before any AI call runs.

✓

Runs as the Real User

Write operations execute under the authenticated user's own session. The AutoCount SDK enforces that user's existing permissions at execution time.

✓

No New Permission Management

Nothing extra to set up or maintain. If a staff member cannot do it in AutoCount, they cannot do it through WareBot.

✓

Audit Trail on Every Write

Confirmation and commit events stream inline, and each committed operation returns a document number, so every change is traceable.

Why Daxonet Builds This

Not an AI bolt-on. Built by the team that already runs your AutoCount.

WareBot is part of Daxonet AI Frontier, the AI operating layer for Malaysian businesses. It is built by people who have implemented AutoCount and warehouse systems for Malaysian SMEs and manufacturers for over a decade, so it reads your real records, not a generic API.

Warehouse write operations, every one confirmation-gated

Input modes in one chat: text, voice and barcode scan

10+

Years implementing AutoCount & business systems in Malaysia

Hours of menu training a new picker needs to start

Take It With You

Download the WareBot brochure.

The full Daxonet AI Frontier WareBot brochure, every capability, all eight operations, the confirmation flow and the security model, in one PDF you can share with your warehouse and operations teams.

Download Brochure (PDF)

Daxonet AI Frontier · WareBot · No form required

What warehouse teams ask before they switch to WareBot

What is WareBot, and how is it different from the AutoCount warehouse screens we use now?

WareBot is a chat-driven assistant that sits on top of your live AutoCount data through Bosun WMS. Instead of navigating four screens to check stock or opening a module and filling a form to move it, your team types or says what they want, 'How much A100 do we have in Warehouse A?' or 'Move 50 units of A100 from WH-A to WH-B', and WareBot does it. It is not a separate copy of your data and it is not a reporting tool; it reads and writes the same records your AutoCount system of record holds, in real time.

What warehouse operations can WareBot actually perform?

Two read operations run instantly with no confirmation: real-time stock balance by item code (optionally filtered by location, all UOMs returned) and item lookup from any barcode or sub-code. Eight write operations run against live AutoCount data and each returns a document number: create stock transfer, create stock request, create stock take, create stock issue, create stock receive, update customer pick (marks the invoice Picked), update customer pack (marks it Completed) and create credit note. Every write pauses for operator confirmation before it commits.

Can warehouse staff use voice, or do they have to type?

Both. Staff can type or record a voice note in plain language. WareBot transcribes audio server-side through a Whisper-compatible service that accepts m4a, mp3, mp4, mpga, wav, webm and ogg, and returns the transcript with a detected language and confidence score. Pickers and packers can keep their hands on the goods and just talk to the app.

Does WareBot work in Malay and Mandarin, not just English?

Yes. WareBot understands any language the underlying speech and language models support, including English, Malay and Mandarin. A picker can ask in Bahasa Melayu, a stock controller can ask in Mandarin, and WareBot understands warehouse terminology without anyone being retrained on a fixed command set.

How does WareBot stop someone from making a wrong stock change by accident?

Every write operation pauses and sends a confirmation prompt before it executes. The operator sees the full operation, what, how many, from where, to where, and confirms or cancels with one tap. Confirmation tokens expire after five minutes and are consumed immediately on use, so a confirmation cannot be replayed. Only after a confirmed commit does WareBot return the document number via an audit event.

Is it secure, and do we have to manage a new set of permissions?

No new permission management is required. WareBot uses the same token authentication as every other Bosun WMS endpoint, and invalid or expired tokens are rejected before any AI call runs. Write operations execute under the authenticated user's own session, so the AutoCount SDK enforces that user's existing permissions at execution time. A staff member can do through WareBot exactly what they are already allowed to do in AutoCount, no more.

Which AI model does WareBot run on, and can it run on our own servers?

WareBot's AI backend is configurable. It can run on a cloud model such as OpenAI GPT-4o, or on a self-hosted OpenAI-compatible model such as Qwen or LLaMA for businesses that want inference to stay on their own infrastructure. The base URL, model name, API key and the Whisper transcription model are all set through XML config or environment variables, which is part of how Daxonet AI Frontier keeps you off single-vendor lock-in.

How long before a new warehouse worker can use it?

Day one. The whole point of WareBot is that there is nothing to learn, there are no menus to memorise and no screen paths to train on. A new picker who can describe what they need in plain language, in English, Malay or Mandarin, can check stock and start operations immediately, while an experienced worker stops miskeying location codes because they can scan a barcode mid-conversation instead.

Your warehouse team could be running stock by chat next week.

Book a 30-minute WareBot demo against your own AutoCount data. We will run the eight write operations live, show you the confirmation flow, and tell you honestly what it takes to switch your pickers, packers and stock controllers to natural-language warehousing.

Build your own copilots — no code required.

From shop-floor to finance — one connected system.

From experiment to operations — in 90 days.