Methodology
How this record was built
This showcase is an independent research project. It is not an official publication of IMDA, SBF, or any winning company, and it carries no endorsement from them.
Official source: SME AI Impact Awards — IMDA & SBF
Award scope
The SME AI Impact Awards 2026 are the inaugural edition, jointly organised by the Infocomm Media Development Authority (IMDA) and the Singapore Business Federation (SBF). The awards sit under the National AI Impact Programme and support Singapore's National AI Strategy.
The organisers received 290 nominations across 21 industry sectors, recognising 30 standout SMEs across two categories: Innovation Excellence, for companies that developed proprietary AI solutions, and Adoption Excellence, for companies that adopted ready-to-use solutions. Winners were announced at SMEs Go Digital Day on 13 October 2026. Eligibility required at least six months of live use with positive business impact, and shortlisted companies received a site visit conducted by SBF to verify their claims. According to IMDA's Singapore Digital Economy Report 2025, national SME AI adoption surged to 23.4% in 2025 (up from 14.5% in 2024), highlighting the selective standard set by the 30 award recipients.
This project covers all 30 winners. It does not cover shortlisted companies that did not win, and it makes no claim about them.
Library cohorts & provenance
The showcase is organized as an ongoing, evidence-graded library across distinct, transparently labeled cohorts. The evidence-grading system is the unifying brand; the awards provided the launch cohort.
The 30 inaugural recipients of the SME AI Impact Awards 2026, organised by IMDA and SBF. This cohort is sealed and immutable. Nothing is ever added to this group. Permalinks remain permanently at /winners/[slug]/.
Non-award Singapore SMEs that submit primary operational documentation, pass the technical inclusion audit, and agree to the 10-dimension evidence rubric. Hosted under /cases/[slug]/ with explicit provenance badging.
Real-world enterprise systems designed, built, or deployed by Ivan Wong and Maple Commerce. Subject to the strictest version of our evidence rules, with mandatory conflict-of-interest disclosure on every surface, paired before-and-after baselines, and fully published architecture stacks.
Canonical source
The primary authority for the winner list, categories and award citations is the official awards page:
https://smesgodigital.gov.sg/web/sme-ai-awards
- Retrieved
- 18 October 2026
- Page last updated
- 16 October 2026
- Winners published
- 30
Extraction method
The awards page renders its winner profiles client-side, so a plain HTTP request does not return the full record. The page was therefore rendered in a headless Chromium browser and the hydrated DOM was queried directly.
Each winner panel was read through stable CSS classes: company name, one-line impact statement, sector, company profile, three headline metrics, three key challenges, the named AI solution and its components. Category membership was derived from the two winner lists in the DOM: records 1 to 15 belong to Innovation Excellence, and records 16 to 30 to Adoption Excellence.
The full page text was also captured as a cross-check. Validation confirmed 30 distinct companies, 15 in each category, with three challenges, three metrics and at least three solution components for every record. No duplicates or missing fields were found.
The awards page does not publish individual winner URLs. Case-study permalinks in this showcase are generated from company names. They are not official URLs.
Evidence hierarchy
| Tier | Source type | Used |
|---|---|---|
| T0 | Official awards page: winner list, categories, citations (Award Cohort). | All 30 winners |
| T1-SUBMITTED | Primary technical documentation, system logs, and architecture schematics supplied directly by company or implementer and reviewed. | Community & practitioner cases |
| T1 | Company sites, official newsrooms, government publications, organiser releases. | 29 records |
| T2 | Named implementation partners, vendor case studies. | 2 records |
| T3 | Independent press, trade publications, third-party reports. | 6 records |
38 distinct sources are cited across the showcase. A source published by the winning company is a primary source about that company, but it is not independent corroboration, and it is never counted as such.
Evidence status
Stated
A source explicitly names the technology or practice in connection with the implementation.
Corroborated
Two or more sources support the claim, at least one of which is not the winning company.
Inferred
A plausible mechanism derived from the described workflow. Always paired with a written basis.
Winner grading
Grade A — Independently supported
At least one source from a party other than the winner addresses the implementation itself, not merely the company.
Grade B — Canonical source supported
The official citation alone establishes the business challenge, the AI solution and the reported outcomes.
Grade C — Limited public evidence
The official record is too thin to support a full case study.
A grade measures available evidence, not the quality of a company's AI implementation. A Grade B company may run a better system than a Grade A company. It has simply published less about it.
Research coverage scoring
Coverage is computed from the structured record, not assigned by impression. Ten dimensions are scored, each worth one point where present and half a point where partial. The total is expressed as a percentage.
01
Company and industry context
A published company profile and sector are available.
02
Business challenges
The citation names three specific challenges rather than a general claim.
03
Named AI solution
The solution has a name that can be traced.
04
Solution components
At least three described components.
05
Named technology vendors
A vendor is named at any stack layer.
06
Documented workflow
The described components support a workflow diagram of five or more nodes.
07
Quantified outcomes
Three published metrics.
08
Independent corroboration
A source from a party other than the winner addresses the implementation.
09
Governance and controls
The citation describes a review, approval or safety control.
10
Implementation timeline
Two or more dated events can be established.
Coverage is displayed separately from the evidence grade and is never presented as an AI maturity rating. The average across the set is 94%.
Inclusion & verification criteria
To maintain the integrity of the record, every non-award implementation (community and practitioner cases) must clear the same quantitative threshold before inclusion:
- Production tenure: At least six months of continuous, live production operation in Singapore. Prototypes, hackathon demos, and unpaid trials are strictly excluded.
- Quantified business outcomes: At least one measured, verifiable operational metric. Paired before-and-after baselines (e.g. cycle time, error rate, labor hours) are required for full scoring.
- Named AI solution: A traceable product, proprietary engine, or commercial stack layer.
- Verification documentation: Direct submission of primary verification artifacts (system screenshots, workflow logs, architectural schematics, or customer sign-offs) under tier
T1-SUBMITTED. - Human-in-the-loop governance: Disclosed human oversight controls, approval mechanisms, and PDPA compliance considerations.
The Direct Documentation Advantage: Because submitted companies provide actual numbers and architecture diagrams rather than brief press citations, community and practitioner cases can achieve 10/10 coverage scores that government citations frequently redact.
Conflict of interest & practitioner disclosure policy
Cases delivered by Ivan Wong or Maple Commerce carry an inherent conflict of interest. To protect the credibility of the entire showcase, practitioner cases are governed by four strict, non-negotiable rules:
- Mandatory three-point disclosure: Every practitioner case must display the verbatim statement on its directory card, page header, and source list:“Implementation delivered by Maple Commerce. This case is written by the implementer; the conflict of interest is disclosed.”
- Setting the documentation benchmark: While public award citations often omit vendors and model choices, practitioner cases must be the best-documented records on the site: publishing exact foundation models, orchestration frameworks, paired before-and-after metrics, and verified workflow diagrams.
- Written client consent & honest grading: Client consent is secured prior to publication. Where confidentiality prevents naming the company, an anonymized archetype is published (e.g. “A 14-staff Precision Engineering Distributor”) and penalized 0.5 points on the identity dimension so the scoring remains mathematically honest.
- Conservative outcome attribution: Practitioner metrics remain strictly categorized as “company-reported”. Outcomes are never upgraded to “independently verified” merely because Maple Commerce engineered the build.
How inference is handled
Inference appears only in a labelled section titled "Unconfirmed — technical inference". Every inference states the known evidence, the inferred mechanism, why the mechanism is plausible, the feasible alternatives, and what remains unverified.
No commercial vendor is inferred without evidence. No performance metric, cost saving or adoption timeline is inferred at all. Where no technical detail exists beyond the award citation, the case says so and offers no inference.
Two diagrams appear on each case study. The documented diagram contains only components the evidence supports. The illustrative diagram is a reference build for other SMEs and is labelled as such.
Research limits
The awards were announced five days before this research was completed. Independent press coverage of individual winners is sparse, and general web search returned a high proportion of irrelevant results for small-company queries. Where a company had no discoverable public material, the case study rests on the official citation and says so.
No winner was contacted for this project. No company was asked to confirm or correct its case study. Figures are reported as published, and company-reported statistics are labelled as such throughout.
About the researcher

Founder of Maple Commerce and BestCourses.sg, with over 20 years of experience in digital marketing, enterprise operations, and automation workflows.
Former Google Partners Trainer and adjunct trainer/lecturer in marketing analytics and AI adoption across institutions including SIT, SMU, and BCG-RISE.
Corrections and attribution
Company names, award citations and reported figures belong to the winning companies and to IMDA and SBF. This project reproduces the substance of the award citations to support verification, and publishes original analysis around them.
No IMDA, SBF, or company logos are used. No affiliation, endorsement or sponsorship is implied. Homepage screenshots are captured at a fixed viewport for identification and comment, dated on the page, linked to the live site, and never presented as the award-winning implementation. The award attribution line used throughout is:
Winner, SME AI Impact Awards 2026 (IMDA & SBF)
To report an error, identify the case study and the specific claim, and cite the source that contradicts it. Corrections are applied to the dataset and the affected case study, with the update date changed.