Brian Arfi Portfolio: A Product Manager Case Study on AI Workflows

Many people type "Brian Arfi" or "Faridhi" into Google to find my portfolio. You might be one of them.

Usually, people look for the long list of companies where I have worked. Tokopedia. Hijra Bank. Flip. Three startups I built from the ground up. Today, I serve as a Product Director leading four product teams for a superapp in the Middle East.

On paper, that resume looks good. But honestly, it is just a list of business entities. It does not show the true market value of a professional.

A product manager's value is not defined by where they sit. It is defined by leverage. It is about how much operational cost they can cut and how efficient the systems they design are. And today, the sharpest tool to create that massive leverage is artificial intelligence.

Efficiency Is an Old Habit

I have an old habit that sticks to every part of my portfolio. I love cleaning up leaking operational costs.

When you see the number of over $4 million in yearly savings on brianarfi.com, you might assume it is the result of modern artificial intelligence. It is not.

Making companies efficient is an old habit of mine, long before AI became a trend: over $4 million per year from various cost-saving initiatives. These included process improvements, infrastructure cost optimization, and sharper product decisions. Automation was just one tool, not the only source. Leverage like that used to require a position, an engineering team, and expensive systems.

Let's look at the facts.

When I led product at Flip, my team and I cut money transfer costs by roughly 32 percent in just six months. That equaled about USD 2.12 million per year. We hit 450 percent of our initial target. We also dropped customer support costs by 42 percent, saving the company around USD 200,000 in seven months. We did this by pushing ticket automation from zero to 70 percent to handle about 10,000 monthly complaints.

Go back further to my Tokopedia days. I led a team of three PMs, 25 engineers, and five designers handling the identity and authentication platform. The traffic was massive. We cut OTP delivery costs by USD 2 million per year while clearing the path for 1.5 million monthly transactions that used to fail. I also led the integration of Tokopedia and Gojek accounts. It was a massive cross-department project. We had a three month target. We finished it in one month through tight execution and strategic planning.

Then, as Head of Product at Hijra Bank, I directed five PMs across four main product streams: acquisition, transaction, identity, and engagement. The Zakat product Gross Transaction Value shot up 549 percent above the normal monthly baseline in just two months. Yearly growth hit 288 percent. I also centralized an authorization platform used by more than 50 collaborators.

Combine those initiatives, and the savings clearly pass $4 million per year.

But there was a bitter reality back then. Saving a company millions of dollars required huge upfront capital. You needed a great, expensive engineering team. You needed robust infrastructure to maintain the systems. The ideas were cheap, but the execution required massive capital.

The Faridhi Portfolio in the AI Workflow Era

Today, reality has completely changed. The principle of making a company efficient remains the same, but the tools are entirely different. Now AI is the sharpest tool for that same habit, and it can be taught to everyone on the team.

Let me show you the living proof from my portfolio in the AI era.

As a Product Director today, I handle the Marketplace, Platform, a B2C Super App, and E-Commerce Solutions for Saudi Arabia and the GCC region. The main focus is unifying all these ecosystems into one complete loyalty super platform. It is a place where users can pay with reward points and businesses can open new monetization doors.

One of the latest initiatives I built is FINA. FINA is an orchestrator for customer support running on an Agentic AI foundation. I built and released the Proof of Concept for FINA in just two weeks. The cost is highly optimized, dropping to roughly USD 0.004 per agent interaction. Even better, before it reached production, FINA caught a data access vulnerability between users. A human eye would have almost certainly missed it.

On the platform side, I designed and released a new seller portal in April 2026. It features bulk merchant catalog onboarding and a complex MGC Legacy Sync architecture. I also rewrote our commercial product strategy. We pivoted from chasing high volume, low yield SMEs to focusing on 16 high yield large tenants using loyalty point burn logic.

That is my main corporate work. Outside of working hours, I build my own personal infrastructure.

In the past, distributing content across many channels required a full operational team. You needed a video editor, a copywriter, and a social media manager.

Now? I built and run an 8-channel content distribution system entirely alone. This AI system automatically cuts my long videos into short clips. It then distributes them to YouTube, Shorts, Reels, Stories, WhatsApp, LinkedIn, and Threads.

I also built the Applied-AI Certification system from scratch. Everything is fully automated. On top of that, I built the AI Circle community. All these AI workflows run consistently in the background. They take inputs, understand context, write drafts, and execute the final output.

Work that used to demand dozens of people and a massive budget is now something I orchestrate alone while drinking coffee in front of my laptop.

Moving Leverage to the Team

Ultimately, this portfolio has one common thread. Your value is not measured by how many lines of code you write. It is not measured by how thick your product documents are.

The true price of a professional is how smartly they place the right tool at the most expensive problem.

Artificial intelligence shifted who gets access to top tier leverage. It used to be a monopoly held by directors or VPs with dozens of engineers. Now, that power belongs to any individual willing to learn. A regular PM, a marketing staff member, or an operations team can now build simple automation scripts using AI. They no longer need to wait months in an IT ticket queue.

This turns my old habit into a culture that can spread massively. AI makes long processes that used to cost millions incredibly cheap.

If you are an individual who wants to experience firsthand how to build AI workflows to increase your daily value and productivity, you can join and learn with me at AI Circle. We will discuss practical things you can use immediately, with no empty theories.

But if you represent a company that needs to transform how your entire team works to achieve that million dollar efficiency, we can discuss it further on the Corporate page.

The most valid portfolio is not a long resume full of adjectives. The most valid portfolio is real results that change how a business runs.