SAN FRANCISCO — The Philippines has built a sizeable pool of Salesforce professionals through its outsourcing industry, but much of the higher-level architecture work remains overseas, according to Cloud Jedi Solutions founder Joey Chan.
Chan, one of only three Salesforce MVPs in Southeast Asia, said Filipino developers and other technology professionals should move beyond implementing specifications and take a larger role in deciding how enterprise systems are designed — an ability that is becoming more important with the rise of agentic artificial intelligence.
“The challenge right now for the Philippines is that a lot of the architecture work is being done onshore, meaning whether it’s US, Australia, or UK. But, the work in terms of execution is being done in the Philippines. It doesn’t have to be that way,” Chan said in an interview during Dreamforce 2026.
Chan, 38, has worked with Salesforce for about 18 years and now leads a local community for Salesforce architects. His company, Cloud Jedi Solutions, helps businesses integrate Salesforce with their existing systems, serving mostly US-based medium-sized companies as well as Philippine clients such as First Gen and BMW.
His entry into the Salesforce ecosystem began while he was studying at the University of Asia and the Pacific, where a Philippine company introduced the platform to students.
After graduating, he joined Professional Services On-Demand as a Salesforce developer before turning to freelance work nine months later and eventually putting up Cloud Jedi.
The company now has fewer than 10 employees, reflecting a small but specialized model that has allowed it to serve clients outside the Philippines.
Chan said the country’s pool of Salesforce professionals has benefited from the onshore-offshore model and the presence of large technology and outsourcing companies such as Accenture.
Cebu has also emerged as another center of Salesforce talent outside Metro Manila, although other regions have yet to develop communities of comparable size.
“I’d say numbers-wise, we’re actually doing fairly well within Southeast Asia. A big part of it is the BPO sector,” he said.
But numbers alone may not be enough as AI changes the skills companies require. Chan said software development is becoming faster with AI-assisted tools, shifting the challenge from writing code to determining what should be built, why it is needed, and how it should fit into an organization’s existing systems.
“I’m biased here, but I’d say the skills to be able to architect these well is going to be much more in demand because, in a way, architecture is something quite broad, but at the same time, you need your foundations to be very solid,” he said.
“Producing code isa lot faster now, but now the question becomes: What do you build and why, and how it actually blends together with what you already have,” he said.
The shift could create an opportunity for Filipino technology workers to take on higher-value roles, but it also exposes a persistent problem among Philippine companies: fragmented systems and disconnected data.
Chan said adopting the core Salesforce platform is generally manageable. The more difficult task is integrating it with legacy applications and deciding how information should move across different parts of the business.
“It’s really when they need to work with all of the different existing systems; how it actually will blend in well with the existing systems,” he said.
The problem becomes more pronounced when companies attempt to deploy agentic AI. AI systems require access to accurate, properly structured, and relevant information, but many enterprises continue to keep data in separate platforms that do not communicate with one another.
“First, you need someone who understands where data needs to reside, what’s the purpose of each system, and how data flows within each system,” Chan said.
“Because you’ve probably heard garbage in, garbage out, right? If you don’t have proper data structures in terms of how you can properly get the right data, it won’t be successful.”
He said data quality has become even more important as enterprises introduce large language models and autonomous AI agents into their operations.
“AI can only work so much. It needs the proper context in terms of working properly,” he said.
Despite the attention surrounding generative and agentic AI, Chan said most Philippine enterprises remain in the experimentation stage. Some organizations have introduced narrowly defined applications, but broader deployment is constrained by cost, technical complexity, and the difficulty of identifying use cases that can produce a measurable return.
“There’s still a lot of experimentation. There are already some niche use cases that I’ve seen that work well,” he said. “I’d say we’re still in the early stages, especially for the enterprise, because there’s a cost.”
For small and medium enterprises, Chan said AI is accessible enough for initial testing, but businesses should resist adopting it merely because the technology is popular.
“The question is finding the proper use cases that would prove a good ROI,” he said. “The question is: Can they really solve their existing problems with AI? Of course, not everything, but you have to be able to understand when it makes sense to use.”
Companies must also determine whether a problem requires AI or can be solved more reliably through conventional software, such as predefined code and automated workflows.
Chan said AI could take over repetitive tasks such as responding to routine inquiries, allowing workers to focus on building automations and handling more complex work. But this transition would require employees to learn new tools and develop skills that generate greater business value.
“In terms of how it’s changing, I’d say it’s really switching from something menial,” he said. “Instead of them just answering questions, they could be the one building a lot of the other automations.”
Cloud Jedi is also experimenting with AI to support its own service delivery. Chan said the company would continue focusing on Salesforce’s core platform, including Sales Cloud, Service Cloud and Experience Cloud, while expanding its work involving system integration.
“Connecting more systems is something much more integral now, much more important because in order for you to be able to use AI, you need to have the data sets there,” he said.
“And if they’re not connected, if they’re silos, then you won’t be able to get AI to do its work.”
