Start with the status of the national framework
DOST introduced NAIS Ph as a proposal in its May 2025 account of National Innovation Day. A later DOST report, published in December, states that the President approved NAIS-PH in May 2025. Reading only the earlier announcement would leave the strategy’s status out of date.[1][2]
This guide examines the implementation pipeline rather than offering another catalogue of national AI ambitions. The factual statements are attributed to official documents; the operating recommendations are Kooya’s analysis. Evidence reviewed for this article does not establish that every agency has adopted the same tools, budgets or delivery timetable.
Separate the framework from its work packages
DOST describes five connected strategies: infrastructure, workforce, innovation, ethics and policy, and deployment, phased through 2028. Its announcement names several departments, education bodies and the Philippine Statistics Authority as participants in the broader ecosystem.[1]
Kooya’s interpretation is that integration cannot be reduced to purchasing a model. An agency might have trained staff but no authorized data pipeline; another might have a working prototype but no operations team. These are different delivery problems and need different owners.
For a proposed project, map those dependencies explicitly. Record who provides the data, who approves its use, who develops the system, who tests it and who receives incidents after launch. A national framework supplies direction; the project still needs an executable division of work.
Do not collapse DOST and DTI into one initiative
DTI launched its National AI Strategy Roadmap 2.0 and Center for AI Research on 3 July 2024. The roadmap builds on a 2021 initiative and addresses research, workforce, data access, connectivity, ethics and adoption. Its documented origin differs from DOST’s later NAIS-PH framework.[5]
Neither an overlapping topic nor a similar acronym proves that one program replaced another. The reviewed documents do not establish a merger or a single intake process. A company should therefore name the program it wants to engage, rather than assume every AI announcement points to the same service.
Kooya recommends explaining the request in ordinary operational language before selecting an institutional label: a research partnership, access to infrastructure, staff development or assistance evaluating an application. That makes the first conversation easier to route and reduces the chance of preparing the wrong proposal.
DICT adds an infrastructure agenda through 2033
The Department of Information and Communications Technology reports that it launched the Philippine AI+ Infrastructure Masterplan 2026–2033 with ADB on September 8, 2026, following presidential approval. Its announcement identifies connectivity, AI computing, data centers, sustainable energy, skills and policy as priorities. The announced capacity and reskilling figures are targets, not completed results.[6]
This gives readers another named initiative to track alongside DOST’s framework and DTI’s roadmap. Kooya’s interpretation is that national integration needs both the research pipeline and the infrastructure that allows applications to operate. The reviewed announcement does not establish that the three programs share one budget, a single application process or identical responsibilities. Keep the program name, accountable office and dated evidence attached to each claim.
Read the research roadmap as a portfolio
PCIEERD’s AI roadmap, updated in September 2025, combines facilities, human resources, policy and mission-driven research. Its project tables distinguish completed, ongoing and terminated work. Its 2026 milestone row includes AI for government operations and service delivery.[4]
The distinction matters. A milestone names an intended result; a status table records a particular reporting snapshot. Neither alone demonstrates that a service is available to a new user or that a completed research project has an operating budget.
Kooya recommends maintaining a simple evidence record for each opportunity: the target, the latest dated project update, the delivery organization and the next public checkpoint. If the material stops at the roadmap, describe the opportunity as a research priority. If a pilot is reported, retain that qualification until an operating service is documented.
Look at funding documents for the next implementation step
PCIEERD’s January 2026 call document is explicitly for 2028 funding. It corroborates NAIS-PH approval and an October 2025 stakeholders’ conference. Its AI priorities include locally relevant applications, cybersecurity, emerging AI approaches and Philippine-language capability.[3]
A future funding call is useful evidence of intended direction, but it does not show that a particular application received support. It also should not be advertised as open today without checking the application window and any subsequent notices.
Kooya’s advice is to distinguish the year of the call, the intended funding year and the proposed implementation period in any internal brief. A supplier or university team planning staffing around an assumed immediate start can otherwise build its schedule on the wrong date.
Our DOST funding guide separates research, commercialization and financing pathways, including the limits of the checked application windows.
Make a proposed use case specific enough to evaluate
Consider a fictional agency that spends hours routing incoming requests. A useful research question would be whether a system can suggest the correct destination from an approved set, explain its evidence and defer ambiguous submissions to an officer. “Use AI for government services” is too broad to test.
Kooya recommends specifying the current process, the proposed assistance and the boundaries in one page. Include representative examples, an agreed reference answer and the human decision that remains necessary. Evaluate incorrect routing and missed urgency separately from average response time.
The same discipline helps a research partner decide whether the challenge actually requires new research, adaptation of an existing method or a straightforward software improvement. It also makes the project easier to explain to the eventual service owner.
Plan the handover before the prototype is celebrated
A prototype can establish technical feasibility while leaving integration unresolved. Kooya recommends agreeing early who will maintain connectors, review errors, approve model changes and pay recurring costs. The receiving organization should participate in testing, not first encounter the system at a closing demonstration.
Write down what must happen when the system cannot process a request or the source data changes. Keep the original workflow available during a bounded trial. Establish which results need human approval and what evidence would justify expanding access.
These are implementation recommendations, not claims that DOST requires this exact checklist. Their purpose is to make the transition from research output to an accountable service visible, especially where project funding and everyday operations belong to different teams.
Judge progress by the next verifiable delivery decision
For a prospective participant, the most useful follow-up is concrete: which documented program matches the problem, what stage it has reached, and what evidence the next decision requires? That produces a more actionable brief than combining every announced AI investment into a single rollout story.
Kooya would track problem ownership, authorized data, evaluation results, operator readiness and the next decision date. Where information is unavailable, leave the field unresolved rather than infer success or failure. A strategy can be approved while individual services remain at different stages; a careful implementation brief preserves that distinction.