October update: Caraga reports training for 380 MSMEs
In a report published on October 15, 2026, DOST-Caraga says 380 MSMEs across its five provinces underwent practical AI training from July to September under Project AIM. The first rollout covered four cities and three municipalities. Sessions led by AI Master Trainers addressed content creation, marketing and business management. A second batch was scheduled for the remaining months of 2026.[1]
This adds evidence of delivered training in another region. The report does not provide measured savings, sustained-use rates or a count of production AI deployments. The second batch is a stated plan, not evidence that enrollment is currently open. Prospective participants should confirm dates, eligibility and availability with the relevant DOST office.
Kooya’s recommended follow-up is a short adoption record: the task selected, the person responsible, the approved information used, the output checked and whether the workflow remains useful after the session. For example, an entrepreneur testing product descriptions could track factual corrections and review time before publishing. This is a proposed measurement method, not a reported Project AIM result.
A regional plan makes the national ambition concrete
DOST’s regional AI work offers a useful view of how integration might happen outside national laboratories. Western Visayas launched its AI Development Action Plan in February 2025. Its nine areas connect talent, research, industry, infrastructure, data access, startups, partnerships, governance and public awareness. This is a broader agenda than teaching people to use a chatbot.[2]
The practical question is whether those pieces become a working service for a school, local government or business. A regional strategy can create access to expertise and equipment, but a local organization still needs to identify the task, assign responsibility and judge the result. The implementation framework below is Kooya’s analysis of that last mile, not a DOST requirement or a report of commercial outcomes.
Computing investment is an input, not a deployment count
In January 2026, DOST VI reported funding turnover and pre-implementation work for regional high-performance computing facilities and a smart-industry initiative. Five HPC facilities were planned at four universities and DOST Western Visayas. The same announcement described support to fully operationalize ISAT U’s SMARTECH Center. Its wording does not establish that every facility is already operational.[3]
For a potential user, the next questions are service questions: who can request access, what training is required, which workloads are supported and who maintains the environment? Those answers matter more than a theoretical computing specification. An organization should also check whether its first use case actually requires substantial computing; a basic process improvement may depend more on cleaner records and a reliable integration.
For teams preparing an access inquiry, our infrastructure guide distinguishes compute, connectivity and model-reuse services.
Business outreach provides an entry point
The May 2026 AI Fest launch outlined activities covering research, demonstrations, creative work and policy. Its official AI Connect page scheduled an August 4 business forum organized by PCCI–Iloilo for entrepreneurs, executives and MSMEs, with demonstrations and business matching. These are evidence of an adoption-oriented program design. They are not evidence of a specific productivity gain. The advertised event date has passed.[5][6]
After an event, a business needs a smaller next step than “transform the company.” Take one recurring bottleneck and write down who performs it, what information they need, how long it takes and where mistakes occur. That produces a useful conversation with a researcher or implementation partner and prevents a compelling demonstration from becoming an unrelated software purchase.
Treat workforce capability as part of the workflow
Kooya’s recommendation is to train people against the work they will actually own. A supervisor needs to recognize an unacceptable output and stop a process. An operator needs to correct incomplete inputs and escalate exceptions. A technical maintainer needs to understand access, monitoring and recovery. A general awareness session can introduce these responsibilities, but it cannot demonstrate competence in all of them.
Give each role a short practice exercise using representative cases. Include at least one situation where the correct action is to pause. Observe whether the team can identify the issue, explain it and complete the manual alternative. This makes training measurable without turning it into a contest of prompt-writing tricks. Retain the exercise so a replacement employee can be trained later.
DICT’s roadshows show another regional route
PIA’s July 30, 2026 report describes DICT Region 2’s two-day Cagayan Valley AI Roadshow with Google Philippines and Cagayan State University. It brought together public-sector and education participants for practical AI sessions. The account illustrates skills outreach in a different region; it does not establish a formal joint delivery program with DOST Western Visayas or measured productivity gains.[8]
Read together, the initiatives suggest several possible entry points: a research partnership, regional facilities, business engagement or staff training. A local team should choose the route that matches its problem. In Kooya’s view, attendance becomes useful operational capability only when participants can apply the learning to a defined task, recognize an error and sustain the process afterward.
Governance has to include the people affected
A November 2025 regional workshop produced an initial draft of Western Visayas AI governance and ethics guidelines; its report discussed privacy, bias, transparency, accountability and workforce concerns. Separately, DEPDev VI reported an August 5, 2026 AI Policy Lab involving marginalized sectors. Consultation and drafting are meaningful steps, but neither source proves that a binding final regional AI rulebook has been enacted.[4][7]
For a local pilot, participation should have a practical consequence. Ask frontline users what a wrong answer would cost them, which decisions they need explained and how they would appeal an outcome. Use those answers to choose review points and escalation paths. A system serving the public should not quietly remove an accessible human route simply because a digital route is available.
Measure a complete task, including correction
Consider a hypothetical food processor evaluating an AI demand forecast. The relevant result is not the attractiveness of the forecast screen. Compare the proposed purchasing decisions with a baseline, including stockouts, spoilage and the time the purchasing officer spends reviewing suggestions. Preserve the officer’s approval responsibility during the trial. This example is an illustration, not a reported DOST-funded deployment.
For a service team, the equivalent measures might be completed requests, correction time and successful handoffs. Review results by task type and operating conditions. A strong average can conceal failure on a smaller but important group of requests. Include the cost of subscriptions, integration, review and support so a claimed saving survives a complete comparison.
Make the route from training to adoption visible
A useful regional adoption dashboard separates inputs from outcomes. Participants trained, facilities funded and events delivered describe activity. Workflows still in use after a pilot, operators able to handle exceptions and measured improvements describe adoption. Neither category replaces the other. Reporting both helps teams identify whether a promising initiative is stuck at awareness, access, validation or maintenance.
Each pilot should have a named workflow owner and a short review date. Record the baseline, the agreed success threshold and the failure conditions before testing. When a project does not proceed, preserve the reason. An honest decision to stop an unsuitable application is valuable evidence for the next team, especially when the same data limitation could recur across several institutions.
The next regional milestone is repeatable use
Western Visayas provides a concrete regional example of DOST-linked planning, infrastructure preparation and engagement. The sources support that direction; they do not establish universal access, completed installation of every planned facility or a quantified regional return on AI. Those distinctions keep the conversation focused on what implementation still needs to prove.
For a Philippine organization, the immediate action is a one-page pilot brief: one problem, one owner, the data required, the current baseline, a reviewable test and a maintenance plan. Bring that brief to the relevant regional office or prospective research partner and ask which existing service fits. Regional ambition becomes useful when local teams can turn it into work they understand, measure and sustain.