Hands-on AI adoption for busy teams

Your people build AI tools for their own work, in the tools you already pay for. We work alongside them until AI is part of the day job.

Kuala Lumpur, working across Malaysia and Singapore

Who we are

Amok Collective helps organisations get practical value from the AI tools they already have. We work directly with employees to find the recurring tasks that take up their time, help them build and start using an AI tool for that task in the organisation's own approved tools, and show what changed.

We bring a method, not software: staff keep what they build and know how to repeat it.

Why clients work with us

  • Hands-on. Staff build alongside us, so the tools stay in use after we leave.
  • Tool-neutral. Claude, Microsoft 365 Copilot, ChatGPT, Gemini and in-house GPTs.
  • Practical. We start from real recurring work and show what changed.
  • Small and senior. The founders do the work themselves.
  • Operators, not IT sales. Our backgrounds are commodity and derivatives trading, and AI skilling at national scale. We run our own businesses with AI agents every day.

What we do

01

AI adoption pilot

A small group of staff each build and start using an AI tool for their own recurring work. Short preparation, a few days on site, a written report and a review a month later.

02

Build and automate

Assistants, agents and simple automations for a defined workflow. Larger integrations are designed with your IT or BA team and handed over to them.

03

AI readiness and training

Where to start, what is safe to share, and practical sessions that teach teams to build with the tools they already have.

How a pilot runs

Typically 4 to 6 participants from 2 to 3 functions. Scope and timing are agreed with each client.

  1. About 2 weeks, remote

    Prepare

    Briefing with your BA or IT team, a short survey and a call with each participant.

    You get one focus task per person, agreed with the sponsor.

  2. 4 to 5 days

    On site

    We sit alongside staff, with daily build sessions, go-live and a debrief.

    You get a working tool per participant, in daily use.

  3. About 1 week

    Write up

    A short write-up of each tool and a prioritised list of further opportunities.

    You get one consolidated report with recommendations.

  4. About 30 days later

    Review

    A repeat survey, a usage check and a review call.

    You get before-and-after results.

What staff typically build

Illustrative. The real list comes from the preparation phase with each participant.

FunctionRecurring taskWhat gets built
Trading and researchWeekly market commentaryA project holding the house style, past notes and sources that drafts the first version
Finance and analyticsMonthly Excel analysis and variance notesA saved routine or short script that runs the steps and drafts the commentary
Operations and contractsChecking confirmations against standard termsA clause checklist that returns a table of differences for review
ComplianceThe same policy questions, asked again and againAn agent on the policy library that answers with the source document
Any functionMeetings and routine emailRecaps, action lists and draft replies from standard wording

We did it in our own businesses first

At Diamond Oil and Fats, where Mankarpal is Chief Strategy Officer

Live

AI agents working every day across multiple areas of the business

Repeatable

The method packaged into a kit that other departments used to deploy their own agents

In-house

Operational tools built with AI by the business itself, not by an IT vendor

At Mereka, where Gurpreet led the AI skilling business

250k+

Learners reached by digital and AI skilling programmes

RM2m

AI skilling revenue built from zero in 12 months

5 hrs

Saved per person per week after an internal AI roll-out

The founders

Equal partners who do the work themselves.

Mankarpal Singh Deo

Mankarpal Singh Deo

Co-founder and Director. Trading, finance and operations

Chief Strategy Officer at Diamond Oil and Fats, directing trading and long-term sales to global oil majors. Previously Head of Trading and Business Planning at FatHopes Energy (2019 to 2024) and Vice President, Derivatives Trading at AmBank (2014 to 2019).

Has designed and deployed AI agents across multiple areas of his organisation, and turned the method into a step-by-step kit other departments have used. Builds operational tools and AI products end to end with AI coding tools.

BSc MORSE, University of Warwick · CQF, Wilmott · MIT Professional Education, Applied Agentic AI for Organizational Transformation

LinkedIn
Gurpreet Singh Dhillon

Gurpreet Singh Dhillon

Co-founder and Director. Building and leading organisations

Director, Applied AI and Ventures at Mereka (Chief Executive Officer 2021 to 2026), where he built the AI skilling business from zero to about RM2 million revenue in 12 months. Co-founder of Biji-Biji Initiative.

Has led programmes with Microsoft, IBM, Google.org, UNDP and MDEC, rolled out AI across a team of 22, and designs agentic systems for document evaluation, governed publishing and grant research.

BSc (Hons) Physics, University of Manchester · US State Department International Visitor Leadership Program

LinkedIn

What you can expect

  • Working tools in daily use, not slideware
  • Success measures agreed with you before we start
  • Evidence of what changed: usage, quality, time or confidence, whichever matters to you
  • A team that can maintain the tools and repeat the approach without us
  • An honest view of what did not work

Your data stays yours

  • We use only the tools and data you approve
  • Nothing leaves your systems
  • We follow your security and IT policies, under an NDA signed before access
  • What your staff build belongs to you

Getting started

  1. Pick a team and two or three recurring tasks
  2. Agree the tools and data rules
  3. Book a two-week preparation phase

Talk to us

Mankarpal Singh Deomsdeo@amokcollective.com
+60 12 936 6604

Gurpreet Singh Dhillongurpreet@amokcollective.com
+60 12 214 5573