| THE LOOKOUT | Singapore · 11 Oct 2026 |
Reviewed 100 sources across 2 evidence bands this cycle — 3 developments mattered: the 3 forces below.
Bank teller · monthly brief
Several moderate signals point the same way: routine enquiries, information retrieval and separable workflows face pressure, while broader finance-training pipelines add competition around entry-level banking work. Direct AI adoption in teller work has not yet been observed, so this is a watch-item rather than a settled outcome. Rules provide a partial counterweight: tighter controls around credit systems and higher foreign-hiring thresholds can preserve value in compliant, locally grounded work. For someone in this role, the more likely path is a narrower job centred on exceptions, judgment and accountable customer service. The solid read is pressure on routine work, with regulated human checkpoints retaining value; more is stirring, and the Lookout is watching.
Your role at a glance
| ▼ | How the work itself is changing | Pressures the role |
| ▼ | Competition for your seat | Pressures the role |
| ▲ | Rules and regulation | Supports the role |
+4 smaller forces moving — nothing material.
The lead
▼ How the work itself is changing
Pressures the role
DBS reports that Joy now handles substantial volumes of common corporate-banking requests, showing how service queries can move away from staff. Vending-Bench gains and Wells Fargo’s announced branch-banker agents add context that retrieval and routine workflows are becoming easier to automate, while also giving staff faster support. This points toward fluency with tools on exposed tasks and deeper judgment where a person remains accountable.
Additional sources — googlecloudpresscorner.com · ocbc.com · mindstudio.ai · sg.news.yahoo.com · openai.com · anthropic.com · quoreka.com · leanware.co · fintech.global · anthropic.com (2)
Also moving
IBF’s heavy funding for CFA Level I widens access to finance training while also signalling planned demand, and other programmes continue to supply work-ready entrants. For teller work, that makes differentiation through a scarcer mix of customer, digital and banking skills the stronger position.
cfasocietysingapore.org · acra.gov.sg · ntu.edu.sg
Additional sources — britcham.org.sg · mom.gov.sg · theaccountant-online.com · supportgowhere.life.gov.sg
EU AI Act obligations now cover high-risk credit-scoring systems, while MOM’s higher financial-services pass floors raise some foreign-hiring costs. That can support local, accountable roles around customer checks and exceptions, pointing toward work that carries the compliance load as rules tighten.
squirepattonboggs.com · mom.gov.sg
Additional sources — wsg.gov.sg · allenandgledhill.com · envoyglobal.com · rajahtannasia.com · rajahtannasia.com (2) · mas.gov.sg · ibf.org.sg · mas.gov.sg (2) · mas.gov.sg (3) · csa.gov.sg
The wider market
New firm formation
Not shown: formation figures are published per industry cluster, and none has been matched to you.
Employer hiring build-out
No qualifying build-outs stored for this occupation this cycle.
Counts for the entire MCF snapshot — every employer and occupation — not for the rows above. Postings considered: 114379 · Postings excluded: 74984 (employer 790, occupation 47355, credibility 26839)
MCF coverage is forced by the Fair Consideration Framework rather than organic, so two groups are structurally under-represented. Senior roles exempt from the FCF advertising requirement appear at roughly 1 in 400 postings, so the senior end — finance especially — is largely absent. Micro fund managers and family offices hire through recruiters or networks and never post under their own UEN (0 of 16 observed), so their demand is invisible here.
Newly observed in the MAS register
No register periods stored yet — the figure is a difference between two published traversals, so it appears once a second one lands.
AI benchmarks that test work like yours
Each observed score above is read from Epoch AI's aggregator exactly as published, never re-graded or re-weighted — who originally published it is disclosed per result, below. The 12-month change is the Lookout's own comparison between two such observations; the scores being compared are still exactly as published. These are published test results, so they are a floor: a model given better tools or more time can do better.
A benchmark exercises activities that appear in your role; it does not measure your job.
Evidence bands this cycle: “In the reporting” — independent press and analyst coverage · “From the source” — institutions speaking for themselves.
Every claim above links to an independent source (not every wider-market figure links to an independent source: the platform's own computed statistics, any benchmark result that is reported by a model's own vendor or whose origin is not established, and which of your role's activities a benchmark exercises — the platform's own curated match, not the benchmark publisher's — are marked as such above). The Lookout always shows both sides: the risks to this job, and the new chances it's gaining. · hello@thelookout.ai
This page includes information from the O*NET 30.2 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the CC BY 4.0 license. O*NET® is a trademark of USDOL/ETA. The Lookout (Manithan, Inc.) has modified all or some of this information. USDOL/ETA has not approved, endorsed, or tested these modifications.
This is the public edition of the Brief for this role, built from government and established sources. Get the Brief for your own role, every month, by email.
Get my Brief →