Muhammad Luthfi Arifin
Track D · 2026
01 / 16
Think → Analyze → Build

Build what
people want.
Learn faster
with AI.

A founder’s story about knowing what to build.
M. Luthfi Arifin · Product Engineer · Founder
The product
02 / 16

It worked.

  • Tests pass
  • App shipped
  • Payments live
Everything an engineer can fix
Meme: happy about zero crashes, then shocked by zero downloads
The market
Orders today
0
Customers

So why wouldn’t it sell?

A working product ≠ a wanted product.

Act I · WIGO, 2021–2024
03 / 16
01
What WIGO taught me

How to build.

A B2B and B2C tourism ecosystem, built from zero.

WIGO tourism app: ticket QR and destination discovery screens
Act I · WIGO
04 / 16
We knew
how to build.
I co-founded WIGO and built its tourism ecosystem from zero. Mobile · Web · B2B · B2C · Payments · Analytics
For travelers
B2C

Tickets, wallet, discovery.

For attractions
B2B

Partner platform and dashboards.

The catch
≠

The problem was outside the product.

The real problem
05 / 16
Meme: skeleton on a bench, app developers waiting for users after doing zero marketing
Where our customers came from

Our network
wasn’t the market.

AOur network
Someone knows someone. An intro, a demo, some interest… and we kept building.
BThe market
Hundreds of businesses we never asked what they actually needed.

We asked “Where can we sell this?” instead of “What does this market need?”

Act II · Abadikan
06 / 16
02
What Abadikan taught me

What to build.

Sell early, listen, change, repeat.

Act II · Abadikan
07 / 16
About six months of R&D, in the open

This time, the market
joined the R&D.

V1 → Market
“Not enough.”
People didn’t care. Good. That’s data.
V2 → Market
“Interesting, but…”
They liked it, but wouldn’t pay. Also data.
V3 → Market
People pay.
And they asked for something different.
Build → Market → Learn → Change ↺
Act II · Abadikan
08 / 16
Customers started giving us the answer

We stopped pushing.
The market started pulling.

Abadikan.id digital invitation on three phones
B2C · Abadikan.idDigital invitations → event experiences
Interactive hotel catalog and resort estate map on phones
B2B · “Can you build this for us?”Static PDFs → interactive sales experiences

At WIGO, we tried to find a market for the product.
At Abadikan, we let the market shape the product.

The actual lesson
09 / 16
Changing the first question
Engineers ask
“Can we build it?”
Business asks first
“Should we build it?”
01

Think.

Whose problem is this, and how badly do they need it solved?

02

Analyze.

What does the evidence from the market say?

03

Build.

Only after understanding the problem.

Act III · AI
10 / 16
03
What AI helps me do

Learn faster.

Shorten the distance between a market signal and a business decision.

The bottleneck moved
11 / 16

Building
is fast.

  • Frameworks & cloud
  • Copilots
  • Coding agents
Faster every month
The market
Signal → decision

Learning
is slow.

BeforeSignalDecision
With AISignalAIDecision

Build 10× faster but learn at the same speed, and you build the wrong thing 10× faster.

Act III · AI
12 / 16
Where AI fits

AI doesn’t replace the loop.
It makes the loop faster.

01ThinkWhose problem, and how badly?
02AnalyzeWhat are the signals telling us? AI
03BuildThe next step the evidence supports.
04MarketSell it, ship it, listen.
05DataOrders, questions, complaints, leads. AI
ThinkAnalyzeBuildMarketData
AI
Signals in, decisions faster
Act III · AI
13 / 16
How I use AI today

AI where the business is slow.

Abadikan Metabase dashboard: revenue versus ad spend, numbers blurred
01 · AnalyticsWhat changed, and why?
WhatsApp chat: a customer asks about revisions, Mindi explains, and the customer asks for the order link
02 · Customer supportWhat keeps coming up?
— 03 / SalesWhere are customers stuck?
— 04 / ComplianceWhat needs attention?

The goal isn’t to put AI everywhere. It’s to find where the business is slow.

One important limit
14 / 16

AI finds
signals.

DataPatternsSpeed
You
Still the decision maker

You provide
judgment.

ContextTrade-offsJudgment
15 / 16
Closing

Don’t build faster.
Learn faster.

Think → Analyze → Build → Learn → Repeat
M. Luthfi Arifin · Product Engineer · Founder
laam.my.id
What each taught me
03
01
WIGO

How to build.

02
Abadikan

What to build.

03
AI

How to learn faster.

Build what people want.
Learn faster with AI.

Q&A
16 / 16
Your turn
What are you
building?
And what evidence says people actually want it?
QR code to laam.my.id/talks/build-what-people-want: slides and contact
Slides + connect laam.my.id/talks/build-what-people-want
Meme: Morpheus, nobody makes it the first time
Thank you.
M. Luthfi Arifin laam.my.id