Microsoft Ends Claude Code: How Japanese Firms in the Philippines Should Choose AI Dev Tools

Microsoft Ends Its "Claude Code" License and Shifts to Copilot — Lessons for Japanese Companies in the Philippines Choosing AI Dev Tools

Using a major IT company's pivot in AI-tool strategy as a case study, learn how to choose tools at a Philippine development base without depending on any single tool, along with practical steps that are useful on the ground.

This guide is for Japanese companies considering expansion into the Philippines and for Japanese businesspeople already working there. Using a major IT company's pivot in AI-tool strategy as a case study, we learn how to think about choosing development tools and AI services locally.

---

Part 1: Why This Matters

Step 1: The Philippine Business Context (3 min)

Because English is widely spoken and IT talent is abundant, many Japanese companies locate software development and maintenance bases in the Philippines. In development teams in Manila and Cebu, AI coding-assistance tools are already in daily use. This topic — "which AI tool to choose" — is a question directly tied to the productivity and cost of the local team.

When you rely entirely on a particular overseas tool, there is a danger that the local team's work grinds to a halt if the provider abruptly ends the service or changes the pricing for its own reasons. A tool chosen by Japan headquarters may also not suit the Philippines' connectivity or budget realities. That is exactly why it is important to understand the "dynamics" behind tool selection.

Scene: In a Manila office, you call out to the team leader of your Filipino engineers like this. "The AI tool our dev team uses — we never know when the provider's policy might change it, right? Let's read today's article together so we're prepared not to be in trouble even if there's a change."

Step 2: Key Points From the Original Article (5 min)

We drew out only the facts from the original article and summarized the key points in a table.

Source: Forbes — "Microsoft Ends Claude Code Licenses As It Shifts Developers To Copilot" (June 1, 2026)

This table was created for learning purposes based on facts from public information. For details, please check the original article at the link above.

Related: see How Customizable AI Tool Integration Helps Philippine SMEs Streamline Operations.

Step 3: Comprehension Check (5 min)

Try answering the following questions.

Q1. Whose tool — from which company — is Microsoft ending its license for, after having used it until now? (Hint: Look at the "Event" row in the table.)

Q2. What is Microsoft trying to put at the center in place of Claude Code? (Hint: It's something it developed in-house.)

Q3. When did Anthropic release its new tool, and over roughly what period did users migrate afterward? (Hint: The release was in May 2025. For the period, focus on the figure "nine months.")

Q4. In one large survey, what proportion of respondents named Claude Code as their favorite? (Hint: The phrase "about half" is the point.)

Q5. Does Microsoft believe that "what decides the market is model performance," or that it's a different factor? (Hint: The phrase "distribution power" is the clue.)

---

Related: see How AI Training Helps Philippine SMEs Build Practical Workforce Skills.

Part 2: Putting It Into Practice

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

What we can learn from this topic is the importance of "preparing not to depend too much on any one AI tool." Here are the steps for putting this thinking into practice at a Philippine development base, summarized in a table.

When thinking about budget, you can reduce waste by not making a large investment from the start, but testing with a few people before expanding. Because it also touches local accounting and tax, consult your local accountant early so that the method of booking the costs conforms to the rules of the BIR (Bureau of Internal Revenue).

Step 5: Common Failures and Countermeasures (5 min)

Here are three failures that easily occur when introducing AI tools in the Philippines.

Failure pattern 1: "Relying entirely on a single tool"

Bad example: You put one overseas tool alone into every team and prepared no alternative for when the provider changes its policy. As a result, development stopped when the service was discontinued.

Good example: You decide on one main tool while also testing one alternative that can handle the same work. You keep yourself in a state where you can switch over within a few days even if the provider changes its policy.

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

Bad example: You input source code entrusted by a customer into an overseas tool with the setting that uses it for training. You later found it might run afoul of the data privacy law and scrambled to respond.

Good example: Before you start using it, you check whether the setting prevents the input code from being used for training. If you handle customer data, you also sort out whether consent has been obtained, in line with the guidelines of the NPC (National Privacy Commission).

Failure pattern 3: "Forcing Japan-headquarters standards alone onto the local site"

Bad example: You introduced, as-is, a high-spec tool decided in Japan, but it was too heavy for the local connection to use, and local staff ended up not using it.

Good example: After conveying Japan headquarters' policy, you verify together with local leaders whether it fits the local connectivity and budget. If it doesn't fit, you choose a lighter alternative together.

---

Part 3: Going Deeper

Step 6: Related Technical Terms (5 min)

Claude Code (Anthropic's AI coding-assistance tool) is a tool by which, when a person describes in words the program they want to write, the AI writes or fixes the code for them. At Philippine development bases, you can see junior engineers leaving the first draft to Claude Code and spending their time on review and revision, thereby shortening delivery times.

GitHub Copilot (AI coding assistance offered by GitHub) is, so to speak, the code version of "predictive text," suggesting the next line as you write a program. It is used as a common foundation when a Manila team works in the same environment as developers worldwide.

CLI (command-line interface) is a way of operating a computer by typing command lines on the keyboard rather than using buttons or icons. As with Copilot CLI, situations where local engineers accustomed to such operation use it to work faster are increasing.

A large language model (LLM) is the central mechanism of an AI: it has read huge volumes of text and learned the patterns of language and code. It is also running behind the scenes when a Philippine customer-service team automatically drafts replies to inquiries.

Distribution power refers to the strength of already being embedded where many people work, separate from the merits of the product itself. The original article explains that Microsoft prioritizes this "distribution power" over performance, teaching us that "ease of continued use" also matters when choosing a tool.

Step 7: Thinking About Applying It to Your Company (10 min)

Make visible how dependent you are on the AI tools you use

Write out which work would stop if a tool your local team currently uses suddenly became unavailable.

Prompt: Using "could we keep operations running for a week without this tool?" as a standard, the strength of your dependence becomes visible.

Next action: Together with local leaders, list the AI tools in use and rate the impact of each going down on a three-level scale.

Which to prioritize: "performance" or "ease of continued use"

Consider, in your own situation, whether to choose the top-performing tool or a tool that is somewhat lower in performance but can be used stably over the long term.

Prompt: At the development frontline, performance matters; for core operations used over the long term, stability matters. Think about it by use case.

Next action: Divide your operations into "those needing the latest performance" and "those where stability matters," and summarize on a single sheet the tool policy that suits each.

Preparing for a switchover in case the provider changes its policy

So you won't be in trouble even if a tool's provider changes its policy, practice moving to an alternative.

Prompt: An alternative you've never tried is useless when the moment comes. Testing it small is your insurance.

Next action: Next month, set aside a half-day for the local team to try an alternative tool and get a feel for switching over.

---

Part 4: FAQ

Q1. In the end, which AI tool should a Philippine development team choose? There is no single right answer. What matters is to decide on a main tool while also preparing an alternative. Because connectivity varies by region in the Philippines, it is safest to decide only after actually testing on the ground.

Q2. Is it OK to input a customer's source code into an overseas AI tool? First, check whether the setting prevents the input code from being used for training. The Philippines has a data privacy law, and the NPC (National Privacy Commission) oversees its operation. If you handle customer data, also sort out how to obtain consent.

Q3. What should I do when a tool decided by Japan headquarters doesn't fit the local site? The shortcut is to respect headquarters' policy while showing, with concrete numbers, whether it fits the local connectivity and budget. Because verbal agreements easily diverge later in the Philippines, always put the agreed content in writing.

Q4. How should I budget for the tool costs? Grasp the monthly cost in both peso and yen terms, and prepare for exchange-rate movements as well. Because the accounting treatment of costs also touches tax, consult your local accountant early so it conforms to BIR (Bureau of Internal Revenue) rules, to be safe.

Q5. Is news of a major company switching tools relevant to small and mid-sized Japanese companies too? Very much so. The moves of major companies can be a precursor to a shift in industry-wide standards. Regularly check what policy the provider of the tool you use is taking, and put in place a structure to notice changes early.

---

Tips for Making It Work (3 Tips)

List the AI tools you currently use, and write out the impact if they go down. When you can see how much you rely on which tool, you can respond calmly even to a sudden policy change. Doing this together with local leaders also reflects the frontline's real sense.

In addition to your main tool, test one alternative. An alternative you've never used is useless when the moment comes. Even a half-day of testing greatly reduces the anxiety of switching over.

Separate your tool-selection criteria into "performance" and "ease of continued use." Prioritizing performance for frontline development and stability for operations used over the long term prevents wasteful switching. Summarizing the policy by use case on a single sheet keeps your decisions from wavering.

---

Bonus: How to Use PH AI Works

PH AI Works supports the use of AI and technology in the Philippines. In connection with this theme, we provide practical support tailored to local circumstances.

As a next step, you can consult us on matters such as the following.

• We sort out the AI tools your local development team uses and check, together, the degree of dependence and ease of switching.
• Regarding the handling of customer data and source code, we sort out an operating method in line with the Philippines' data privacy law.
• So you won't be in trouble even if the provider changes its policy, we help you build a switchover procedure to an alternative.

Please feel free to contact us. Consultations are free.

---

References and Sources

• Forbes — "Microsoft Ends Claude Code Licenses As It Shifts Developers To Copilot" (June 1, 2026)
• National Privacy Commission (NPC)
• Bureau of Internal Revenue (BIR)