▸ Retail head walkthrough

Opportunity cockpit

Loan growth is a deployment problem · not a demand problem

The customer base already holds the signals. A trigger acted on within 48 hours converts at several times the rate of one acted on after three weeks — the engine makes that visible.

Addressable opportunity

OMR 184m

in existing base · modelled drawdown

Customers with live trigger

14,842

today · unactioned 8,310

Triggers fired · last 30d

24,180

conversion rate 17.4%

Incremental drawdown

OMR 42.1m

last 30d · avg time trigger→decision 2.3d

Opportunity by product

Personal loan

Addressable · modelled conv.

OMR 68m

8,420 cust · 22% modelled

Housing loan

Pre-approvals

OMR 42m

1,208 cust · 18% modelled

Auto loan

Vehicle intent markers

OMR 24m

4,100 cust · 24% modelled

Card limit increase

Utilisation signals

OMR 18m

6,200 cust · 31% modelled

Top-up on existing

Debt-free + life events

OMR 20m

3,180 cust · 26% modelled

Buyout of external debt

Competitor wallet share

OMR 12m

12,840 cust · 14% modelled

Trigger funnel · last 30 days

Fired

24,180

Qualified by credit policy

18,860

Offer made

12,340

Customer engaged

6,760

Application started

5,120

Approved

4,580

Drawn

4,210

Leakage point: largest drop between offer made → customer engaged (−45%). Causes rank: wrong channel (31%), timing (24%), framing (18%). Opportunity: Journey Builder & channel optimiser.

Aging decay · conversion by days since trigger

Why time-to-action matters

A trigger acted on within 48 hours converts at 4.2× the rate of one acted on after 3 weeks.

Channel performance · cost per acquisition

Last 30 days

Channel	Fired	Converted	Rate	Cost / acq.
SMS	8,420	984	11.7%	OMR 3
App push	6,200	1,340	21.6%	OMR 2
Internet banking msg	3,840	720	18.8%	OMR 2
Outbound call	2,120	612	28.9%	OMR 24
Branch conversation	1,840	672	36.5%	OMR 42
Relationship manager	1,760	812	46.1%	OMR 86

Trigger Studio

Compose · backtest · activate

A business user can build a new trigger live, in under two minutes. Live backtest on every edit.

Rule composer

Signal

Condition

Threshold · OMR 500

Qualifying filter

AND/OR combine with…

Suppression window · 90 days

Offer

Channel

Would have fired

14,200

last 24m backtest

Modelled conversion

18.4%

base rate

Expected drawdown

OMR 47m

over 12m

Live backtest · last 24 months

ran in 892ms

Backtest window: 2024-08 to 2026-08 · signal: first_school_fee_payment · n=14,200 fires · 2,614 converted to education finance within 90d · modelled conversion 18.4% ± 1.2pp

Sensitivity · threshold moves

Threshold	Fired	Modelled conv.	Expected OMR	Signal quality

Overlap with existing triggers

Prioritisation rule: expected value · if the customer fires this and another together, this takes precedence.

Trigger catalogue · 24 triggers across 5 families

Every trigger detectable from data the bank already holds

Each entry shows the data sources, inference confidence, customer-benefit case, suppression rules, and false-positive rate from backtesting.

Family A · Competitor wallet share Highest-value family in the catalogue

#	Trigger	Detection logic	Offer	Addressable	Conv.	FP rate	Benefit
A1	External loan maturing	Recurring debit to named bank, consistent amount, declining residual count inferred	Buyout + top-up	12,840	14%	3.1%	cust better off
A2	External loan instalment identified	Monthly SO matching loan-instalment characteristics	Refinance at better rate	18,420	8%	5.8%	cust better off
A3	External credit card settlement	Monthly payment to another bank's card entity	Balance transfer	8,200	11%	2.4%	cust better off
A4	External mortgage detected	Large recurring debit to bank or housing fund, long duration	Mortgage refinance	4,100	6%	4.2%	bank-only
A5	Salary split	Part of salary immediately transferred to another bank	Full salary transfer + loan incentive	6,800	22%	1.8%	mutual
A6	Competitor deposit build	Regular outward transfer to another bank's savings	Deposit consolidation + lending headroom	3,420	18%	7.2%	bank-only

Family B · Capacity & life events

#	Trigger	Detection logic	Offer	Addressable	Conv.	FP rate
B1	Internal loan maturing	Bank Muscat loan with 1-3 instalments remaining	Top-up or new facility	4,200	42%	1.1%
B2	Salary increment	Salary credit rises 5%+ and sustains two months	Limit increase or new facility	8,400	14%	3.8%
B3	Promotion step-change	Salary credit rises 15%+ in a single month	Upgrade of product tier + facility	2,100	26%	2.4%
B4	Bonus or variable credit	Non-salary credit from the employer	Investment, prepayment or deposit	14,200	11%	4.1%
B5	End-of-service payment	Large one-off credit consistent with gratuity	Deposit and advisory, not lending	1,840	38%	1.4%
B6	New dependant	Beneficiary added, or school-fee spend appears	Education finance	6,840	22%	5.2%
B7	Marriage indicators	Joint account opened, cluster of furnishing spend	Personal loan	1,840	31%	3.1%
B8	Spouse salary arrives	Second salary credit to the household address	Joint affordability reassessment, mortgage	1,120	41%	1.8%
B9	Debt-free status reached	Last instalment of all external and internal debt paid	Full-capacity new facility	3,840	52%	0.8%

Family C · Purchase intent

#	Trigger	Addressable	Conv.
C1	Vehicle shopping	4,100	24%
C2	Property shopping	2,840	18%
C3	Home improvement	3,200	16%
C4	Travel booking	8,420	12%
C5	Education enrolment	5,100	28%
C6	Medical expenditure	2,840	22%
C7	Large single purchase	12,400	14%

Family D · Product gap & cross-sell

#	Trigger	Addressable	Conv.
D1	No credit card	38,200	16%
D2	Card limit saturation	8,400	22%
D3	Overdraft dependence	2,840	34%
D4	Cash-advance reliance cust saves	1,420	42%
D5	Islamic preference signal	6,100	18%
D6	Single-product customer	18,200	—

Family E · Retention & defence

#	Trigger	Detection logic	Offer	Addressable	Conv.
E1	Settlement enquiry	Customer requests early settlement quote	Retention offer before refi away	840	38%
E2	Salary diversion risk	Salary credit stops/reduces without resignation markers	Urgent RM contact	410	22%
E3	Competitor onboarding	New outward transfer to another bank's account in customer's own name	Retention contact	2,100	14%
E4	Dormancy onset	Transaction frequency falls below threshold	Reactivation (not lending)	8,400	11%
E5	Maturity without rollover	Term deposit matures and funds leave	Win-back offer	1,840	24%
E6	Complaint followed by inactivity	Complaint logged, engagement drops	Relationship recovery · lending suppressed	120	—

Customer opportunity card

For a two-minute branch conversation

Not an analytics screen. One sentence of plain language, evidence, and three possible actions.

Salim Al Hinai

Teacher · Ministry of Education · Nizwa

Civil ••••0342 · Customer 9y 2m · Products: 3 · Rel. value OMR 4,800

Trigger A1 · Live

Salary day: 25th · Stability: 0.97

"Salim's auto loan with Bank Dhofar matures in two months. His instalment is OMR 182. He has held his salary with us for nine years."

Why now · evidence

Standing order of OMR 182 to Bank Dhofar	28th of every month
Series duration	46 months
Inferred: 5y auto loan, 14 instalments remaining	matures Dec 2027
Original amount (inferred)	~OMR 9,400
Residual balance (inferred)	~OMR 2,548

The offer

Product	Buyout + top-up
Buyout residual	OMR 2,548
New personal loan	OMR 12,000
Total facility	OMR 14,548
Tenure	60 months
Indicative rate	6.75%
Monthly instalment	OMR 231
vs current OMR 182 at Bank Dhofar + future borrowing elsewhere	cust better off net

Affordability · at a glance

Current debt burden ratio	15.4%
Post-offer DBR	19.6%
Policy ceiling (salaried)	50%
Headroom	30.4 pp

Live from Prototype 3's affordability engine · open in Credit Decisioning

Talking points · generated for this customer

Opening

"Mr Al Hinai, I noticed your Bank Dhofar auto instalment of OMR 182 is close to ending. Would you like to look at bringing it to us with better terms?"

Likely objection

"But I'll be paying more per month." Response: "This includes a OMR 12k top-up — the comparable instalment on just the buyout would be OMR 49."

"If you can show two months' Bank Dhofar statements, we can finalise today. Shall I prepare the paperwork?"

Actions

Branch daily list · Ruwi

18 contacts for tomorrow

Ranked by expected value, capped at a workable number per staff member. Walk-in match on card swipe.

Opportunities to call

capped from 42 live triggers

Expected drawdown

OMR 142k

if all converted

Walk-in queue

live match · action now

#	Customer	Trigger	Offer	Value	Conv.	Channel	Assigned to
1	Salim Al Hinai	A1 · External loan maturing	Buyout + top-up · OMR 14,548	OMR 12,000	31%	Branch	Maryam A.
2	Fatma Al Zadjali	B2 · Salary increment	Limit top-up · OMR 4,500	OMR 4,500	24%	App push	auto
3	Said Al Harthy	C1 · Vehicle shopping	Pre-approved auto · OMR 14,000	OMR 14,000	28%	Branch	Omar M.
4	Noor Al Rawahi	E1 · Settlement enquiry	Retention offer · OMR 18,000	OMR 18,000	19%	Call	Branch mgr
5	Khalid Al Shukaili	B1 · Internal loan maturing	Top-up · OMR 8,000	OMR 8,000	22%	Branch	Maryam A.
6	Buthaina Al Maskari	B9 · Debt-free reached	Pre-approved OMR 45,000	OMR 45,000	18%	RM	Rel. mgr
7	Hamad Al Ghafri	D3 · Overdraft dep.	Restructure to term loan · OMR 3,500	OMR 3,500	34%	Branch	Omar M.
8	Shamsa Al Harthy	C5 · School enrolment	Education finance · OMR 2,400	OMR 2,400	26%	App push	auto
9	Majid Al Hashmi	D1 · No card	Card at OMR 2,500 limit	OMR 2,500	16%	SMS	auto
10	Thuraya Al Kindi	B5 · End-of-service	Deposit + advisory	OMR 32,000	38%	RM	Rel. mgr
11	Nasser Al Lawati	A2 · Ext loan instalment	Refinance · OMR 6,800	OMR 6,800	14%	Call	Branch mgr
12	Jokha Al Riyami	C3 · Home improvement	Home improvement · OMR 5,000	OMR 5,000	16%	App push	auto

Team view · yesterday's performance

Staff	Contacts	Offers	Conversions	Value	Vs plan
Maryam Al Lawati	14	11	4	OMR 28,000	+18%
Omar Al Hinai	12	9	3	OMR 22,000	on plan
Nawal Al Rawahi (RM)	6	5	3	OMR 48,000	+32%
Branch manager (walk-in)	4	3	1	OMR 8,000	−12%

Journey builder

Multi-step journeys · Trigger A1 example

External loan maturing. Journey adapts to engagement at each step.

Journey · "External loan maturing · buyout"

DAY 1

SMS

"Your Bank Dhofar instalment ends soon. We have an offer."

28% engage

DAY 3

App push

"OMR 14,548 pre-approved. See terms."

22% engage (of remainder)

DAY 7

Branch call

RM-led conversation if still unengaged

41% convert

DAY 14

Suppress

If still unengaged, mark offer dormant, log for 90-day retry

Journey performance · last 90 days

Step	Sent	Engaged	Converted	Keep?
Day 1 SMS	1,840	516	142	keep
Day 3 App push	1,324	291	98	keep
Day 7 Branch call	1,033	422	174	keep
Day 14 Supplementary SMS	611	18	3	drop · fatigue cost

Language variants · respecting customer preference

Each message exists in Arabic and English. Delivery picks the customer's declared language preference. For customers with no declared preference, language inferred from Arabic-script usage in app and branch activity. Measured: 72% Arabic, 28% English in the sample base.

Day 1 Arabic variant

قسطك الشهري لبنك ظفار على وشك الانتهاء. لدينا عرض لك.

Risk guardrails

Where the engine declines to pursue

Every trigger passes credit policy before an offer is made, not after the customer engages. This panel surfaces the volume of opportunities the engine declined to pursue.

Triggers suppressed · this month

847

cases the engine declined to pursue

Vulnerability stress markers

312

late/short salary · overdraft excess · card advance

Debt-burden ceiling breach

248

would exceed 50% / 60% policy limit

In-arrears / IFRS9 stage 2+

142

suppressed as hard-exclude

Exclusion register · drivers

Vulnerability stress markers

312

DBR would breach ceiling

248

IFRS 9 stage 2 or 3

142

Complaint flag active

Asalah · RM-controlled

Vulnerability markers · explicit

Salary credit arriving late or short in 2 of last 6 months
Card cash-advance usage
Overdraft excess days ≥ 5 in a month
Standing order failure
Declining end-of-month balance trend over 3 months

Suppresses lending triggers even where policy would allow the loan. Decision captured for audit.

Portfolio drift monitor

Credit quality of trigger-sourced lending vs existing book

Median DBR of trigger-sourced lending 14.2% vs existing book 16.8%. No month-over-month drift. The engine is not loosening standards.

Shared affordability engine · Credit Decisioning

Every trigger-sourced application runs the same affordability engine as a branch-sourced application in Credit Decisioning. The two cannot disagree because they are the same model. Click any customer in the trigger list → opens their case in the Credit Decisioning Case Workbench.

What-if simulator

Trigger engine · scenario builder

Which triggers active · what sensitivity · what contact capacity. Live recomputation under 2s.

Input controls

Triggers active

Trigger sensitivity

Contact capacity · 1,200/week

Offer aggressiveness · −25 bps rate

Contact fatigue cap · 2/mo

Credit policy stance

Horizon

Incremental drawdown · 12m

OMR 184m

net new book

Net interest income

OMR 11.2m

net of funding + ECL + acq. cost

IFRS 9 stage-1 ECL at origination

OMR 1.52m

from Credit Decisioning

Customer-impact ledger

OMR 2.14m saved

6,420 cust better off

Contact capacity vs demand

Demand exceeds capacity in weeks 7-12. Prioritisation applies: ranked by expected value. 6,840 lower-ranked triggers deferred — tracked in queue; some will decay.

Cannibalisation estimate

~22% of this incremental volume would have been booked anyway through existing branch and call-centre channels. Net new to Bank Muscat: ~OMR 143m. Finance will ask this question — the answer is pre-computed.

Forecasting

Trigger exhaustion curve

The stock of customers carrying an unactioned trigger depletes as the engine works through it — first-year yields exceed the steady-state run rate.

Unactioned trigger stock

Year-one depletion · steady-state run rate

Stock at launch: 14,842 unactioned triggers · year-one drawdown ~RO 184m · steady-state ~RO 86m/year as new triggers fire at the natural event rate. First-year yields exceed steady state.

Trigger volume forecast · by family

Monthly, with seasonality

Seasonality encoded

Jan · school fees

B6 and C5 triggers up 2.4× monthly avg. Education finance peak.

Eid · bonus credits

B4 trigger spike. Investment and deposit offers prioritised over new credit.

Jul-Aug · summer travel

C4 trigger up 3.1×. Travel instalment plans fire predictably on airline aggregator patterns.

Khareef · Dhofar

Dhofar C3/C7 spike June-September. Visitor spend blurs resident intent — suppression applied.

Pre-Ramadan

C7 large single purchases peak 2w before. Auto triggers (C1) up 1.6×.

Nov-Dec · property handover

Sultan Haitham scheduled handovers drive C2 and B8 clusters.

Backtest

Metric	Model (Jul25-Jul26)	Actual	MAE
Trigger fires · monthly	22,400 avg	24,100 avg	7.2%
Conversion rate · overall	17.1%	17.4%	0.3 pp
Drawdown · total	OMR 198m	OMR 210m	5.7%
Average time trigger → decision	2.1 days	2.3 days	8.7%

Over-prediction concentrated in C4 (Travel) — model under-weighted post-pandemic normalisation of travel spending. Addressed in v2.1.

Conversational analyst

Ask a question in natural language and the analyst queries the dataset, runs the trigger and policy models, and answers in numbers with the records traceable in two clicks.

Trigger Lending analyst

Claude · tool access to Trigger, Policy, Suppression engines

Ready. I can query triggers by family or branch, draft a new trigger rule and backtest it live, decompose any funnel number, show the suppression register, and refuse to decline a case the system should not decline.

Models: TRIG-v2.1 · POLICY-v2.14 · SUPP-v1.4 · 24 trigger definitions loaded

Shared platform · one customer across all three apps

From any trigger card…

→ open the same customer in Credit Decisioning to see the credit decision
→ open in Deposit Pricing to see the funding cost of this facility
→ the three screens show the same affordability number because they call the same engine

Automated trigger discovery

The system runs continuously against the dataset to find behavioural patterns that precede borrowing but are not yet encoded as triggers. 3 candidate patterns awaiting human review this month. Never auto-activated.

Current candidates

"Insurance renewal + 10% vehicle service spend" → auto loan precursor
"Repeated OMR 50-100 debits to utility aggregator" → cross-sell signal
"Al Mazyona prize-draw entry during Eid week" → savings behaviour change