Identical banking queries posed to different AI platforms result in completely different top brand recommendations in 43% of cases, according to new research from Singapore-based agency PRecious Communications.The study, titled AI Authority Index: Banking and published in September, examined 1,628 matched comparisons across 11 AI service proxies (representing platforms like ChatGPT, Gemini, Perplexity, and Claude) across six Southeast Asian markets: Singapore, Malaysia, Indonesia, the Philippines, Thailand, and Vietnam.The findings reveal a fragmented AI visibility landscape for corporate communications professionals, as brand prominence fluctuates wildly depending on the platform, prompt framing, and language used.Key Findings: Language, Prompts, and Market Dominance• Platform Inconsistency: In 706 of 1,628 matched comparisons (43%), AI platforms named a different bank first despite receiving identical prompts.• Language Disparities: Brand rankings shifted significantly when queries were switched from English to local languages. Thailand saw the largest shift (60% of top-ranked banks changed between English and Thai answers), followed by Indonesia (38%), Vietnam (32%), the Philippines (30%), Malaysia (17%), and Singapore (15%). Local-language answers also cited fewer competitors overall (e.g., Malay responses named 1.36 fewer banks on average than English equivalents).• Inclusion vs. Top Rank Gap: Being mentioned by AI does not guarantee top positioning:o Singapore: DBS dominated both visibility (96.3% inclusion) and top ranking (81.7% first-mention share). While OCBC (87.5%) and UOB (84.2%) were frequently mentioned, they earned first-place mentions in only 5.4% and 6.3% of answers, respectively.o Malaysia: Maybank achieved a 91.7% inclusion rate and a 74.4% first-mention share, compared to CIMB's 78.9% inclusion and 2.2% top-rank share.o Philippines: BDO led in top rankings (50.6% first-mention share) despite a lower overall inclusion rate (76.1%) compared to BPI (82.8% inclusion, but only 16.1% first-mention share).• Category Complexity: Business banking queries prompted AI to recommend more brands (average 4.73 banks) compared to personal banking queries (3.52 banks), trust/selection queries (4.53 banks), and public-role queries (3.88 banks).Strategic Implications for Communications LeadersThe report underscores that traditional PR metrics—which measure mere brand mentions or volume—fail to account for how AI engines position brands against competitors.Echoing recent findings from agency Leoprd showing that higher citation volume doesn't inherently translate to stronger AI recommendations, PRecious Communications emphasized that brand reputation is now routinely shaped before companies have the opportunity to make their case directly to consumers. Regional teams monitoring AI visibility purely in English risk overlooking critical competitive dynamics in local-language AI ecosystems.(Photo by POURIA ???? on Unsplash)