Frontiermind
/
Capability build
Live
01 / 08
Mission
A strategic priority becomes the source material for the platform.
Mission priority detected
Protect customers from scam-related financial harm.
Improve detection, escalation timing, and customer communication on the frontline.
Sources ingested
Scam response SOP
✓
Transfer controls
✓
Vulnerable customer framework
✓
Escalation matrix
✓
Prior incident notes
✓
Workflow named
Suspicious Transfer & Scam Intervention
Workflow standard codified
Suspicious Transfer & Scam Intervention
8steps
5evidence rules
3hard gates
6failure patterns
Identify customer intent
Verify identity
Probe source of instructionGate
Document red flags
Assess vulnerabilityGate
Calibrate risk
Escalate before actionGate
Hand-off to fraud ops
Approved standard
Suspicious Transfer & Scam Intervention
generates
Learning pathway
14 modules
Assessment checks
12 items
Simulation pack
8 scenarios
Nomi coaching
Voice agent
Scoring rubric
20 competencies
Manager view
Live readiness
Live · turn 4 / 8
Mrs Tan
Long-time client
"They told me I have to move the funds today."
Evidence gates
Identity verified
Intent stated
Source unverified
Escalate
AR
Aisha Rahman
Frontline Banking Officer
81%
Readiness
StrengthCalm customer communication
GapEscalation triggered too late
Competency scorecard
16 / 20 met
Scam-risk recognition
92%
Customer de-escalation
89%
Escalation timing
58%
Evidence discipline
64%
Documentation accuracy
83%
+ 15 more competencies
4 gaps to close
Nomi next practice
Updated today
Aisha Rahman
Frontline Banking Officer
Scam-risk detection
Proven
Customer de-escalation
Strong
Evidence discipline
Developing
Escalation timing
Developing
Adjacent paths
Senior Service Officer
68%
Branch Ops Lead
57%
Scam Assist AI
Same workflow. Same evidence rules. Same standard as Aisha.
Canonical
57/60
Edge
21/30
Destructive
16/20
Regression
10/10
Sandbox pass rate
87%
Not approved · Retest on coercion cases
Manager
Suspicious Transfer & Scam Intervention
Human · 28
18 ready
7 dev
3 risk
AI · v0.4
87% pass
Blocked
Dimension
Human
AI
Scam-risk recognition
91%
95%
Escalation timing
58%
62%
Customer comms
89%
84%
Documentation
83%
90%
Action required
Targeted practice on escalation · agent retest on coercion edge cases
Human
—
Passport
—
Agent
—
Manager
—
01 / 08 · auto