What the "Devin-kun" Phenomenon Teaches Us About AI Agents: The Impact on Philippine IT/BPO Work and Japanese Companies

Using "Devin," the AI agent spreading rapidly across Japan, as a lens, we map its impact on Philippine IT and BPO work and lay out the steps to adopt it. A practical, plain-language guide for Japanese companies in the Philippines and those considering entry — covering implementation steps, regulations, and how to avoid common failures.

Author
AuthorAuthor

AI Engineer · 36+ years in IT · Japanese, based in Manila for 13+ years

Learning From the "Devin-kun" Phenomenon: How an AI Software Engineer Is Reshaping Philippine IT/BPO Work and Japanese Companies

Using the rise of "Devin," the AI software engineer everyone in Japan is talking about, as our starting point, we explain the changes AI agents are bringing to Philippine IT and BPO work — and the practical points Japanese companies should prepare for.


Part 1: Why This Matters

Step 1: The Philippine Business Context (3 min)

The source article describes how Japan is embracing "AI agents" surprisingly fast. An AI agent is an AI that, once you give it a task, carries out the work on its own. The star of the article is an AI coding tool called "Devin," built by a company named Cognition AI: give it instructions and it writes the program itself, fixes its own mistakes, and even ships it.

This topic is not a distant matter for Japanese people working in the Philippines. The Philippines is a major hub for BPO (business process outsourcing) — call centers, back-office work, and other tasks handled on behalf of other companies. The source article reports that as AI agents begin doing this kind of work more cheaply, the share prices of India's major IT firms have fallen sharply. Philippine BPO and IT work sits in the path of the same wave of change.

At the same time, this is also a major opportunity. According to the source article, one of Japanese companies' weaknesses is low English ability, which has tended to cut them off from the rest of the world. Because the Philippines has an abundance of English-capable talent, it has the potential to become the bridge that connects Japanese head offices with their Philippine operations through AI.

[Scene] An office in Manila. Morning coffee in hand, you turn to your Japanese colleague and say: "Apparently there's an AI called 'Devin-kun' that everyone in Japan is talking about. You just give it instructions and it writes the program for you. Maybe it's time we rethought how we run the admin and IT work at our Philippine site too." That one remark can be the spark that gets your site rethinking the way it works.

Step 2: The Key Points of the Source Article (5 min)

We have pulled out only the facts stated in the source article and organized them in the table below. The figures and proper nouns use only what appears in the article.

ItemDetails
What Devin isAn AI coding tool developed by Cognition AI (a San Francisco startup founded in 2023). Give it instructions and it writes code on its own, fixes its mistakes, and ships it
Popularity in JapanBy user engagement, Japan was the company's first- or second-most popular country in the world. Japanese users gave it the affectionate nickname "Devin-kun"
Japan's market backdropPeople aged 65 and over make up about 30% of residents. The working-age population is projected to shrink by more than 30% by 2060. In 2023 the Ministry of Economy, Trade and Industry estimated that Japan would be short about 789,000 software engineers by 2030
The Sapporo caseA legacy program of over 1 million lines needed overhauling. Work that would normally take 200 engineer-months was completed with Devin in roughly a quarter of the time
FundraisingIn May 2025 the company raised over $1 billion at a valuation of $26 billion. Its annualized revenue grew from $37 million a year earlier to $492 million
Impact on India's IT giantsThe share prices of Infosys, Wipro, Tata Consultancy Services, and HCLTech fell 30–40% over the past 12 months
U.S. AI firms and JapanOpenAI and Anthropic opened their first overseas offices in Tokyo. Japan's three megabanks (MUFG, Mizuho, and Sumitomo Mitsui) briefly gained access to Anthropic's "Mythos" model

Source: Fortune — "'Devin-kun': Japan embraces agents as legacy code and a shrinking workforce create a perfect market for an AI software engineer" (July 3, 2026)

This table was compiled from facts in publicly available information for educational purposes. For details, please see the original article at the link above.

Related: see How AI Agents Help Philippine SMEs Automate Daily Business Operations.

Step 3: Comprehension Check (5 min)

Here are five questions to check your understanding of the source article. Try to recall the answers as you read on.

Q1. Why did Cognition AI choose Japan as its first step into Asia?

Hint: the article cites three factors — Japan's population, its shrinking working-age generation, and its legacy programs.

Q2. In the Sapporo case, how much was the work time reduced by using Devin?

Hint: recall what fraction of the usual 200 engineer-months the work shrank to.

Q3. In 2023, how large did the Ministry of Economy, Trade and Industry estimate the 2030 software-engineer shortage would be?

Hint: it was a big number — roughly 790,000 people.

Q4. What concern caused the share prices of India's major IT firms to fall?

Hint: the reason is that AI agents can do the same work far more cheaply.

Q5. Why is having geographically distant teams an advantage for AI companies?

Hint: when it is daytime in Japan, it is nighttime in New York. It relates to how compute is used.


Related: see How AI Agent Development Helps Philippine Businesses Automate Beyond Prompt Engineering.

Part 2: Putting It Into Practice

Step 4: Adoption Steps in the Philippines (10 min)

Here is how to approach adopting AI agents at a Philippine site, broken into five stages. Each stage comes with points specific to the Philippines.

StageWhat to doPoints to watch in the Philippines
1. Choose the target workStart with work that is highly manual and repetitiveIt is safest to try it first on internal-facing work such as call centers, shared accounting services, or the IT help desk
2. Set a budgetEstimate a budget in pesos for a small trialStart at a scale of a few tens of thousands of pesos per month and expand as you see results — this reduces the risk of failure
3. Confirm how data is handledSeparate information that may be given to the AI from information that must not beThe Philippines has a law protecting personal information, overseen by the National Privacy Commission (NPC). Confirm how customer information is handled beforehand
4. Explain it to local staffConvey that AI is a tool to help, not a tool to take jobs awayIn the Philippines, there are situations where verbal agreement carries weight. Provide face-to-face explanation, not just documents
5. Measure results and expandRecord the time and cost saved and expand to the next work areaDeciding your success metrics (the numbers that gauge how well you did) up front makes it easier to report to management

In Stage 1, it is important not to widen the scope too much. Start by narrowing to one or two tasks, confirm that they are working well, and then move on. In Stage 3, do not put the legal check off until later. When personal information is involved, it is reassuring to confirm with the vendor before adoption whether the AI can be configured not to use your data for training.

Step 5: Common Failures and Countermeasures (5 min)

Here are three failures that easily occur when adopting AI agents in the Philippines.

Failure pattern 1: Assuming "if we leave it to AI, we won't need people"

An AI agent only delivers its full power when there is someone to give it instructions and someone to check the results. If your only goal is to cut headcount, quality drops and you lose trust.

Bad example: A company slashed its Philippine site's admin team all at once and tried to leave everything to AI. As a result, no one was left who could fix mistakes, and the site could no longer respond to inquiries from the Japanese head office.

Good example: First, have local staff take on the roles of giving instructions to the AI and checking its results. Shift people's work from simple tasks to the work of checking and judgment.

Failure pattern 2: Starting to use it without confirming how data is handled

If you feed customer or employee information to an AI without proper checks, you risk breaking the law. The Philippines has a law protecting personal information, and violations are subject to penalties.

Bad example: A company entered customers' personal information directly into an AI tool and only later realized it "had been used for training." This became a situation requiring a report to the National Privacy Commission (NPC).

Good example: Decide in advance the range of information that may be given to the AI. Choose a setting where personal information is not used for training, and make it possible to keep a record of who entered what and when.

Failure pattern 3: Not addressing local staff's anxieties

If you leave anxieties about a new tool unattended, it stops being used on the ground. Because human connection is valued in the Philippines, a one-way announcement alone invites pushback.

Bad example: A company simply notified its Philippine site by email of an adoption policy decided by the Japanese head office, and left it at that. Staff took it to mean "our jobs are going away," and cooperation could not be obtained.

Good example: Explain the purpose of the adoption, and how staff's roles will change, face-to-face. Always set aside time to take questions, and proceed little by little while reflecting their input.


Part 3: Going Deeper

An AI agent is an AI that, once you give it a task, works out the steps itself and carries the work forward. Without being told every detail, it handles a whole job — writing code, fixing mistakes, and so on. For a Philippine IT support team, you might ask, "Look into this system's error and fix it," and the AI would search out the cause and propose a fix on its own.

Legacy code — programs built long ago that are still running but have become hard to change. Over years of use, the people who built them retire and the underlying design grows outdated, making the code difficult to touch. As with Sapporo in the source article, business systems that Philippine companies have used for decades fall into this category, and situations where AI helps drive their overhaul will become more common.

An AI coding tool is an AI that, like Devin, takes over the work of writing programs itself. Because people can shift to the role of checking the finished product, even a small number of engineers can do a great deal of development. A Philippine development team could leave simple features to the AI and spend its people's time on more complex design and quality checks.

Compute — the computing power an AI needs to run its calculations. The amount available is limited, and it can run short during busy hours. The source article explains that because it is nighttime in New York when it is daytime in Japan, having teams in distant locations lets you route processing to the quieter hours. A Philippine site can likewise take advantage of being in a different time zone from Japan and the U.S.

Multilingual AI — an AI that can understand and use several languages, such as Japanese and English, equally well. The source article notes that a Japanese engineer can use Devin in Japanese while working alongside a team on the other side of the world. In the Philippines, combining English-capable talent with AI makes it easier to bridge Japanese head offices and overseas business partners.

Step 7: Thinking About How to Apply This at Your Company (10 min)

Here are three discussion topics for thinking about how to use this at your own company or team. Each comes with a "prompt to think about" and a "next action."

Which work at your Philippine site should you try AI agents on first?

Prompt to think about: the more manual and repetitive a task, the more easily it delivers results. Choosing internal-facing work first lets you experiment while keeping the impact on the outside small. Rather than jumping straight to customer service, it is safer to build up experience internally.

Next action: List out your site's tasks and pick one that meets both criteria — "highly repetitive" and "mistakes can be fixed internally."

How should you evolve your BPO business for the AI era?

Prompt to think about: in the source article, AI agents doing the same work cheaply pushed down the share prices of India's IT giants. Philippine BPO is in the same position. By shifting people to the side that gives the AI instructions and checks its results, you can move to work that delivers higher value.

Next action: Take your current work and make a list that separates the parts you can leave to AI from the parts only people can handle.

How should you combine strong English ability with AI?

Prompt to think about: the source article points out that Japanese companies tend to be cut off from the wider world by their low English ability. The Philippines has many English-capable people. Combine this strength with AI, and your site can take on the role of bridging Japanese head offices and overseas partners.

Next action: Name one concrete task where your Philippine site could bridge the Japanese head office and overseas partners.


Part 4: FAQ

Q1. Can AI agents like Devin be used in the Philippines too?

The source article describes their spread across Asian countries, and AI coding tools are generally usable in many countries over the internet. Adoption is likely possible in the Philippines as well. However, pricing and terms of use vary by vendor, so always confirm the latest information before signing a contract. Whether the tool your Japanese head office uses can be used as-is at your Philippine site also needs to be checked in advance.

Q2. Will AI agents shrink the Philippine BPO business?

In the source article, AI doing the same work cheaply drove down the share prices of India's IT giants; at the same time, a Cognition executive told the reporter that moving to the side that uses AI makes work more interesting and more impactful. In the Philippines too, reducing simple tasks and shifting to higher-value work — such as giving instructions to the AI and checking quality — will be the key to survival. Rather than assuming decline, it is important to think in terms of changing roles.

Q3. When giving information to an AI, what should you watch out for under Philippine regulations?

The Philippines has a law protecting personal information, overseen by the National Privacy Commission (NPC). Before handing customers' or employees' personal information to an AI, confirm whether that information can be configured not to be used for training. There is much in common with Japan's approach to personal-data protection, but the reporting procedures and points of contact differ from country to country. It is safest to work through the checks together with your counterpart on the Philippine side.

Q4. What advantages does the Philippines, with its strong English ability, have?

The source article cites low English ability as a weakness of Japanese companies. Because the Philippines has an abundance of English-capable talent, it can serve as the bridge connecting Japanese head offices with overseas business partners. Combined with AI's multilingual capability, it also shines in situations such as bridging instructions written in Japanese to exchanges conducted in English. This strength is a major value that domestic sites in Japan do not have.

Q5. How much budget should you set aside for adoption?

Because the source article gives no specific pricing, you will need to confirm figures with the vendor. If you are starting in the Philippines, the safe approach is to first set aside a budget in pesos to try it on a small scale, then expand once you see results. Beyond the tool's usage fees, also factor in the effort of explaining and training local staff. Rather than investing heavily from the start, we recommend scaling up in stages as you confirm results.


Tips for Getting the Most Out of It (3 Tips)

Start small with internal-facing work first Using an AI agent straight away for customer service tends to expose failures publicly. If you begin with internal-facing work such as call centers or shared accounting services, mistakes can be fixed internally when they happen. Build up small wins before expanding.

Shift local staff's roles to "the ones who instruct and check" Treating AI as a headcount-cutting tool drops quality and loses trust. The executive in the source article also says that moving to the side that uses AI makes work more interesting. Have your staff take on the roles of giving instructions to the AI and checking the results.

Decide how data will be handled before adoption Feeding personal information to an AI without checking risks breaking Philippine law. Decide in advance the range of information that may be given to the AI, choose a setting where it is not used for training, and make it possible to keep records. Setting this up at the start takes less effort than fixing it later.


Bonus: How to Work With PH AI Works

PH AI Works supports companies looking to advance their use of AI and technology in the Philippines. When adopting the AI agents that are the theme here, an approach grounded in the Philippine business environment and regulations is essential. We can offer advice tailored to the local setting, including bridging your Japanese head office and Philippine site.

As a next step, you can consult us on matters such as:

  • How to choose which work at your Philippine site to try AI agents on first
  • How to manage data in light of Philippine regulations, including the handling of personal information
  • How to go about explaining it to local staff and rethinking their roles

Please feel free to get in touch. Consultations are free of charge.


References & Sources

About the author

Author
Author

Founder / AI Engineer (36+ years in IT)

  • From Tokyo · based in Manila for 13+ years
  • 36+ years in IT (development, SEO, AI)
  • IBM Certified Generative AI Engineer
  • AI chatbots, RAG & AI agent development

A Japanese AI engineer with 36+ years in IT and 13+ years on the ground in the Philippines. I write from hands-on experience to help Japanese companies adopt AI that actually delivers results — chatbots, workflow automation, AI agents, and AI-driven marketing. Feel free to reach out in Japanese or English.

Free AI Consultation

Tell us your challenges and we'll propose the right AI adoption plan for your business.

Book a Free 30-Minute Consultation

Related Articles

AI Case Study

How West Shore Home Uses AI to Double Revenue with 1,000 Fewer Hires — and Why the "Last Mile" Still Belongs to People

A look at how U.S. remodeler West Shore Home used AI to hold back 1,000 hires while doubling revenue, written for Japanese companies operating in the Philippines. Covers how to roll out automation, guard against hallucinations, explain the change to local staff, and stay mindful of NPC rules.

7/5/2026

AI Case Study

Cisco Rolls Out AI Agents to All 90,000 Employees: A Company-Wide AI Adoption Guide for Japanese Companies in the Philippines

Drawing on Cisco's rollout of AI agents to all 90,000 of its employees, this practical guide walks Japanese companies operating in the Philippines through the adoption steps, cost estimation, NPC regulations, and data handling rules they need to get right.

7/3/2026

AI Case Study

Why Rising Electricity Costs Are Fueling a Rooftop Solar Boom in the Philippines | A Cost Guide for Japanese Companies

A practical guide to the Philippines' surging rooftop solar movement, written for Japanese companies operating locally. It covers Meralco's rate hikes, installation steps, payback periods, and local regulatory pitfalls from a hands-on, cost-management perspective.

6/29/2026

AI Case Study

A 50-Year-Old Shop Turns to TikTok Shop: Retail Digital Transformation Lessons for the Philippines

A look at how a long-standing U.S. snack shop used TikTok Shop and online sales to overcome falling foot traffic, with practical lessons for Japanese companies entering the Philippines — covering how to launch online and social commerce, plus local tax and regulatory points to watch.

6/29/2026

AI Case Study

Protecting Your Company and Talent from Increasingly Sophisticated Job Scams: A Practical Guide for Companies Operating in the Philippines

For HR and recruiting teams at Japanese companies operating in the Philippines: how to spot increasingly sophisticated job scams and recruitment impersonation, and how to defend against them. Covers setting up official hiring channels, alerting staff, and working with the NPC and DOLE.

6/27/2026

AI Case Study

Spotting GEO Scams in the AI Search Era: A Guide to Fake Brand-Mention Services for Japanese Companies in the Philippines

A practical guide to protecting your company from GEO scams in the AI search era. Learn how to spot dubious tactics like PBN placements and fake posts, with contract and procurement tips for Japanese companies operating in the Philippines and Japanese residents on the ground.

6/27/2026