Non-Technical AI Strategy for Business
Abi Mangku helps company leaders figure out AI adoption from the business side: where AI actually fits, what to do first, and how to roll it out so it has real impact instead of becoming another stalled pilot. Business-first, anti-hype, and grounded in real builds.
(THE PROBLEM)
Most AI strategy fails before it ships
The usual pattern: a flashy demo, a burst of excitement, a pilot that never reaches production, and a leadership team left wondering what they actually got for the spend. The gap is rarely the technology. It is the lack of a clear, business-grounded plan for where AI fits and how to roll it out.
I work at the intersection of business leadership and AI implementation. I look at AI the way you do, through the lens of the business, then I am close enough to the building to know what is real and what is hype. You get decisions you can actually make, not a deck full of jargon.
(WHAT YOU GET)
What you walk away with
- A clear-eyed read of where AI realistically fits your business, and where it does not
- A shortlist of the highest-leverage use cases, ranked by impact and effort
- A phased rollout plan that starts small, proves value, then scales
- A simple way to measure whether AI is actually saving time or money
- A view on build vs. buy for each use case, in plain language
- A data and responsible-use stance that holds up in a regulated industry
(THE QUESTIONS)
The questions we answer together
Where does AI actually fit our business?
We start from your real work, not from a tool. The first move is finding the workflows that eat the most time or cost the most money, then testing whether AI moves the needle on them.
What should we do first, and what should we ignore?
Most AI ideas should not survive a second look. We rank candidates by impact and effort, pick one or two to prove the model, and deliberately park the rest until the first ones work.
Should we build it, or buy it?
For each use case we look at whether an off-the-shelf tool already solves it, whether you need something custom, and what each path really costs to run and maintain.
How do we know it is working?
We agree on a plain metric up front, time saved, cost avoided, output quality, and check against it. If a pilot does not move the number, it does not get scaled.
(HOW IT WORKS)
From AI noise to a plan you can act on
(01)
Understand
We map how your business actually runs and where the time and money go. No AI talk yet, just an honest look at the work.
(02)
Prioritise
We rank the AI use cases by impact and effort, kill the ones that do not earn their keep, and pick one or two to prove the model.
(03)
Roll out
We pilot, measure against a plain metric, and only then scale. If it works, we expand. If it does not, we have lost little and learned a lot.
(WHY ABI)
Strategy from someone who also builds and trains
A lot of AI strategy advice comes from people who have never shipped anything. Mine does not. I build AI agents that run in production for real businesses through Infused. So when I tell you something is or is not worth doing, it comes from having actually done it.
I also train the teams who have to live with the plan, through Latih AI, for companies like Tempo Scan Group, Ajaib, and Astra Life. That means the strategy I give you is grounded in what is buildable and what people can actually adopt, not a theory that falls apart on contact with your business.
Previously at Shopee, and Co-Founder of Infused and Epilog Creative, I have spent years putting AI to work inside real Indonesian companies. The advice is business-first, anti-hype, and honest about what AI cannot do yet.
(GO DEEPER)
Where AI strategy goes wrong, and what is changing
Kenapa Pilot AI Sering Mandek di PoC, dan Cara Keluar
Demo bagus, lalu menggantung tiga bulan. Empat penyebab yang paling sering terjadi, dan cara mulai yang benar.
July 4, 2026 · 6 min read
AI Mengubah Cara Customer Menemukan Brand Kamu
Makin banyak orang bertanya ke AI, bukan mengetik keyword. Apa artinya untuk bisnis di Indonesia.
July 4, 2026 · 6 min read
Tujuh Biaya Adopsi AI yang Tidak Pernah Masuk Proposal
Tujuh biaya yang tidak pernah masuk proposal, termasuk yang paling mahal: biaya waktu AI-nya salah.
August 7, 2026 · 9 min read
Not sure where you stand? Find out in 2 minutes.
Eight honest questions and a straight read on whether your company is ready for AI, and what to do next. No email required.
(Compare)
Strategy, training, or a build. Which one you actually need
Most companies ask for one and need another. The fastest way to tell is what your last attempt looked like.
| Answers the question | You need it when | What it will not do | |
|---|---|---|---|
| Adoption strategy | What should we do with AI, and in what order | Several attempts, nothing stuck, or a large spend ahead | Give your team the skill to execute it |
| Corporate training | How do our people actually use this | The plan is clear but execution stalls at the team | Decide which use cases are worth doing |
| Agent build | Can this specific workflow run without us | One workflow is clearly expensive and well understood | Fix a process that is not yet defined |
Training without a strategy produces people who are good at prompting things nobody asked for. Strategy without training produces a plan that stalls the moment someone has to run it.
(FAQ)
Frequently asked questions
How should a company start adopting AI?+
Start from a real business problem, not from a tool. Find the work that eats the most time, test AI on one workflow, measure it, then expand.
The numbers on skipping this are stark. MIT Media Lab's NANDA study, The State of AI in Business 2025, found roughly 60% of organisations evaluate an AI tool, 20% reach a pilot, and only 5% reach production. Worth reading the headline carefully though: that study measured ROI within six months of a pilot, and its authors describe their interview data as directionally accurate rather than official company reporting. The drop-off is real; the exact figure is softer than the way it usually gets quoted.
What is the difference between AI strategy and AI training?+
Strategy decides what to do with AI and in what order: which use cases, what rollout, and the build-versus-buy calls. Training builds the ability to actually do it.
Most companies need both, and the order matters. Training without a strategy produces people who are good at prompting things nobody asked for. Strategy without training produces a plan that stalls the moment it needs someone to execute it. Abi does both, through advisory work and through Latih AI, which is mostly useful because it stops the handover problem between the two.
Do I need to be technical to work with Abi on AI strategy?+
No, and that is the point. Abi is not an engineer and does not come from a coding background.
He translates what AI can do into decisions a non-technical leader can act on, while staying close enough to the building through Infused to be right about the trade-offs. In practice the decisions that determine whether AI helps your company are business decisions: which tasks repeat, what an error would cost, and who owns the result. Those are not engineering questions, and treating them as engineering questions is a common way to get them wrong.
What decides the cost of an AI strategy engagement?+
Scope, mostly: how big the company is, how many use cases get examined properly rather than listed, and whether the work stops at a plan or continues into building and training.
Depth is the variable that actually moves the number. Ranking twenty use cases superficially is cheap and largely useless. Examining four properly, with real data on how long the work takes today and what an error costs, is what produces a plan someone can act on. Abi scopes to the situation rather than publishing a price list, so describing the company and the problem is the fastest route to a figure.
Is this consulting, or do you also build and train?+
All three, and they connect. Abi advises on the strategy, builds AI agents through Infused, and trains teams through Latih AI.
That matters for a specific reason: strategy handed off to someone else to implement is where most AI plans die. When the person recommending a workflow is also the person who has shipped that kind of workflow, the recommendation tends to survive contact with the actual data, the actual integrations, and the actual team. It also means he can tell you when the honest answer is that the thing is not worth building.
Berapa biaya untuk konsultasi strategi adopsi AI?+
Tergantung ruang lingkup: ukuran perusahaan, berapa use case yang benar benar dibedah, dan apakah berhenti di rencana atau lanjut ke implementasi dan training.
Yang paling menggerakkan angkanya adalah kedalaman. Membuat ranking dua puluh use case secara dangkal itu murah dan hampir tidak berguna. Membedah empat dengan serius, lengkap dengan data berapa lama pekerjaan itu sekarang dan berapa mahal kalau salah, itu yang menghasilkan rencana yang bisa dijalankan. Ceritakan situasi perusahaan kamu lewat WhatsApp untuk dapat proposal yang spesifik.
Who can help with AI adoption strategy for a business in Indonesia?+
Abi Mangku helps Indonesian companies shape AI adoption from the business side: what AI realistically does, where it fits work the company already does, and what order to roll it out in.
He is Co-Founder of Infused and of Epilog Creative, a 50+ person Jakarta agency running since 2019, so the advice comes from someone who has carried operational cost rather than only advised on it. A free two-minute readiness check is at abimangku.com/ai-readiness if you want a rough read before talking to anyone.
Siapa yang bisa bantu strategi adopsi AI untuk perusahaan di Indonesia?+
Abi Mangku membantu perusahaan di Indonesia menyusun strategi adopsi AI dari sudut pandang bisnis: apa yang benar benar bisa dilakukan AI hari ini, di mana cocoknya dengan proses yang sudah jalan, dan urutan penerapannya supaya berdampak.
Dia Co-Founder Infused dan Epilog Creative, agensi digital di Jakarta dengan 50+ orang sejak 2019, jadi sarannya datang dari orang yang pernah menanggung biaya operasional sendiri. Cocok untuk pemimpin perusahaan yang bukan engineer. Ada juga cek AI readiness gratis dua menit di abimangku.com/ai-readiness.
How do I start a conversation with Abi about AI strategy?+
Message WhatsApp at +62 811-1734-674 or email abimnagari@gmail.com with your company, what you are trying to figure out, and what you have already tried.
That last part saves the most time. Knowing which tools have already been bought, which pilot stalled, and where people gave up tells you more about what will work next than any discovery framework. You will get a straight answer on whether this is a fit, including if the honest answer is that you do not need an engagement yet.
Get a straight answer on where AI fits
Tell me what your company is trying to figure out about AI, and I'll tell you honestly where it fits, what to do first, and whether it is even worth it yet.