Fc
Fortes Co-pilot
Mission Control · v0.1
Atyab Rehman
IT Lead
AR
Sherlock · Pattern Observation

Intel

What Sherlock has been watching. Eighteen opportunities surfaced this month, each one observed in your actual workflows.

Pattern window:last 90 days·47,212 events·Confidence ≥75%

Opportunities here aren't suggestions from a brochure. They're patterns Sherlock saw repeatedly enough to recommend automation. Each one cites the evidence.

Opportunities discovered · 90d
0
across all 6 Fortes entities
▲ 6 vs previous quarter
Estimated annual value
AED 0.00M
hours reclaimed + revenue lift
confidence-weighted projections
Ready for approval
0
data threshold met · awaiting green-light
Under observation
0
more data needed before recommendation
next threshold check: 7 days
Hero · Decisions waiting on you

Ready for your approval

5 opportunities have crossed the 30-data-point evidence threshold.

🎯

Sibling Re-enrolment Nudge

Pattern observed across Sunmarke + RIS·38 data points·92% confidence
What Sherlock saw

When a family with 2+ children renews the elder sibling, the younger sibling renewal currently lags by an average of 11 days. In 14% of cases the younger sibling's renewal slips past the Term 3 deadline and triggers a late-fee invoice plus a manual chase by the admissions team.

Proposed automation

When elder sibling renewal is confirmed, Watson sends a same-day reminder to the parent specifically about the younger sibling, pre-filled with the younger child's name, year group, and the active sibling discount.

Projected impact
  • 47 fewer late-fee invoices/year
  • AED 12,400 saved in admin time
  • AED 38,000 collected on time vs late (working capital benefit)
Build effort

Low · ~3 days · existing Watson skill extension

Evidence cited
SIMS · 38 sibling-pair renewal sequences (Sep 2025–May 2026)Finance · 11 late-fee invoices auto-flaggedWatson conversation logs · 6 reactive renewal chases
🌐

Russian-Language Inquiry Priority Routing

Pattern observed at Sunmarke·31 data points·87% confidence
What Sherlock saw

Russian-speaking parent inquiries (n=31) convert to tour at 71% — the second-highest of any language segment. But they currently wait an average of 14 minutes for first contact because Watson's Russian quality flag routes them through Sarah Pereira manually.

Proposed automation

Promote Watson's Russian skill from "human-review" to "auto-send" for standard inquiry types (tour booking, fee questions, doc checklist). Continue human review for safeguarding-adjacent or curriculum-deep replies.

Projected impact
  • Russian-language first response: 14 min → <60s
  • Conversion uplift estimate: 71% → 78–82% (modelled on EN/AR uplift when auto-send was enabled)
  • 2–4 additional tour bookings/month
  • Pipeline value: AED 800K–1.2M/year (Sunmarke Russian-speaking LTV)
Build effort

Trivial · ~4 hours · skill permission toggle + 2-week monitor

Risk note

Sample size is modest (n=31). Sherlock recommends a 2-week pilot with daily quality spot-checks by Hassan Mubarak (multilingual lead reviewer).

Evidence cited
Watson logs · 31 Russian inquiries Mar–May 2026Conversion analysis · cross-language CTRResponse time logs · Pereira queue
📚

Year-7 Entry Pattern — JINS Palm Jumeirah Funnel

Pattern observed cross-brand·42 data points·89% confidence
What Sherlock saw

42 families at JINS Palm Jumeirah are entering their pre-school exit year between Sept 2026 and Sept 2027. Of these, 38 live within 8km of Sunmarke JVT and only 4 have inquired about Year 1 placement at either Sunmarke or Regent. The default assumption appears to be that JINS-to-Sunmarke is automatic, but it isn't — applications are still required.

Proposed automation

Pre-emptive Watson conversation series (multilingual) to JINS Palm parents at 14-month, 9-month, and 4-month marks before sibling school start, co-branded with Mrs. Namrata Nayar's signature for warmth.

Projected impact
  • Convert 12–18 additional JINS Palm families to Sunmarke/Regent FS2/Y1
  • Estimated pipeline value (year 1): AED 720K–1.1M
  • Reduces "lost to competitor" leakage at the JINS-to-school transition
Build effort

Medium · ~2 weeks · cross-brand orchestration + Mrs. Nayar voice training

Evidence cited
JINS Palm enrolment data · age cohort analysisAddress proximity analysis (privacy-safe — postcode level only)Historic JINS-to-Sunmarke conversion rate: 41% (vs internal target 65%)
🎤

Parents'-Evening Transcript Mining

Pattern observed at Sunmarke + RIS·34 data points·81% confidence
What Sherlock saw

Parents' evenings produce ~340 individual teacher-parent conversations per school per term. Currently no structured capture. Teachers write follow-up actions by memory after the event. Sherlock detected 19 instances over the last 3 terms where the same parent concern was raised in consecutive parents' evenings without intervention — suggesting follow-ups are slipping.

Proposed automation

Opt-in Lexa transcript (parent consent prompt at booking), automatic summarisation per conversation with action items extracted, routed to the teacher and copied to the Form Tutor for cross-reference.

Projected impact
  • Zero "I told you that last term" parent complaints
  • Teacher post-event admin: 4 hours → 30 min
  • Pastoral pattern detection: cross-teacher visibility
Build effort

High · ~6 weeks · consent workflow + Lexa adaptation + teacher onboarding

Risk note

Requires careful parent consent flow and KHDA data review. Recommend piloting at one year group first (suggest Y10, where parents'-evening stakes are highest).

Evidence cited
Watson + Lexa pattern matcher · 19 repeated concerns flaggedTeacher post-event reflection logs (anonymised)SIMS pastoral notes cross-reference
💰

Late-Payment Soft Tone Experiment

Pattern observed cross-brand·47 data points·94% confidence
What Sherlock saw

Watson currently sends fee reminder messages in a single approved tone (Sunita-approved). 47 reminders sent in the last 90 days resulted in: 32 paid within 48 hours, 11 paid 3–14 days later, 4 ignored entirely and eventually escalated to phone. The 4 ignored cases shared a common signal — they came from families with 2+ children where the elder was a recent leaver. The soft tone might be reading as "easily ignorable" for this segment.

Proposed automation

Branch Watson's fee reminder logic. For "first-attempt + standard family" keep current soft tone. For "second-attempt OR recent-leaver family" switch to a slightly firmer (still polite) variant. Both still gated to Sunita's approval policy.

Projected impact
  • Reduce ignored-reminder rate from 8.5% to ~3%
  • AED 84K freed working capital per year
Build effort

Low · ~1 week · second tone variant + branching rule

Evidence cited
Watson fee-reminder logs · 47 sequencesFinance · 4 escalated-to-phone casesSIMS · recent-leaver flag join
Cooking · Not yet ready

Under observation — 13 patterns gathering evidence

Each pattern auto-promotes to Ready-for-Approval when it crosses 30 data points.

📅

Parents' weekend booking surge

73% of tour booking requests arrive Sat–Sun. Watson responds within 60s, but pre-tour document requests sent Sat–Sun show a 2.4× lower completion rate.

Data points: 22/30
Next threshold: Jun 18, 2026
🚌

Bus-route capacity optimisation

4 morning routes routinely run at 60% capacity. Sherlock's initial read: route consolidation feasible without service impact.

Data points: 18/30
Next threshold: Jul 02, 2026
📊

Staff CV follow-up cadence

HR sends 40–60 candidate rejections monthly. Currently manual templated. Watson could auto-personalise.

Data points: 28/30almost ready
Next threshold: Jun 04, 2026
→ auto-promote to Ready-for-Approval
📝

Year 6 EOY report turnaround variance

Variance of 14 days in report turnaround between Y6 form tutors. Pattern suggests a shared template + Watson assist could compress.

Data points: 24/30
Next threshold: Jun 22, 2026
🎟

JINS waitlist auto-promotion timing

Waitlist conversions cluster Tue–Thu 10–12 GST. Promotion offers sent outside this window convert ~2× lower.

Data points: 19/30
Next threshold: Jul 09, 2026
🀄

Mandarin Academy interest — Korean families

Cluster of 9 Korean-heritage Sunmarke families inquiring about Mandarin Academy enrichment. Currently no Korean-language landing copy.

Data points: 16/30
Next threshold: Jul 18, 2026

Trip permission slip return rate by year group

Y3–Y5 return rate <70% by deadline; Y10+ runs >92%. Single-touch WhatsApp nudge tested in 3 cases lifted return to 88%.

Data points: 21/30
Next threshold: Jun 28, 2026
🧩

Inclusion plan renewal cadence

Inclusion plans set for annual review slip on average 22 days past target. SENDCo team manually chases.

Data points: 17/30
Next threshold: Jul 14, 2026
🎓

Sixth Form open evening attendance vs conversion

Attendance has grown 18% YoY but conversion to application down 6 pts. Watson could test post-event personalised follow-up.

Data points: 25/30almost ready
Next threshold: Jun 15, 2026
→ auto-promote to Ready-for-Approval
🚍

Term-3 transport invoice query pattern

Spike of 38 transport-invoice queries in first 10 days of Term 3. Same questions repeated. FAQ + Watson pre-emptive note could absorb.

Data points: 20/30
Next threshold: Jul 05, 2026
📈

Staff CPD attendance — voluntary vs mandated

Voluntary CPD attendance lifts 41% when invite arrives with personalised relevance reasoning vs flat invite.

Data points: 15/30
Next threshold: Jul 25, 2026
🏆

Sunmarke vs Regent house points engagement

Regent house points engagement 2.1× Sunmarke. Cross-brand best-practice transfer opportunity.

Data points: 22/30
Next threshold: Jun 26, 2026
📖

Library catalogue search intent

23% of library searches return zero results. Sherlock can mine intent to recommend acquisitions.

Data points: 12/30
Next threshold: Aug 02, 2026
Map

Confidence vs impact

Hover any dot for the opportunity name and key numbers.

QUICK WINSSTRATEGIC BETSFILL TIMEAVOIDBUILD EFFORT →ANNUAL IMPACT →
Legend
Ready for approval
Under observation · >75% data threshold
Under observation · <75% data threshold
Sherlock's read

All 5 ready-for-approval opportunities cluster in the Quick Wins quadrant. Combined Phase 1.5 build effort: ~4 weeks. Combined annual value: AED 412K.

Recommend bundling as a "Phase 1.5 expansion" SOW after Phase 1 goes live.

The loop

Recently decided

Humans stay in the loop. Not every Sherlock idea gets greenlit.

Recently approved · now building or live

APPROVED · APR 14
24-hour Uncontacted Lead Alert
Live since May 12 · 0 leads slipped past 48h
Actual value (vs projection): 112%
APPROVED · APR 28
Sunmarke EOY Report Auto-Generation
In build · 73% complete · live Jun 12
Projected value: AED 184K/year saved
APPROVED · MAY 8
JINS Open Day RSVP Bulk Reminder
Live since May 18 · attendance +28% vs prior open day

Recently rejected / deferred (and why)

REJECTED · APR 21
Auto-respond to parent complaints in <30s
Reason: Devika requested human-only handling for emotional context.
Decision logged: parent complaints route to DSL/Form Tutor first, always.
DEFERRED · MAY 2
Cross-school behavioural pattern sharing
Reason: Privacy/RBAC scope unclear. Loop in legal Q3 review.
Re-surface for review: Sep 2026
REJECTED · MAY 14
Predict-then-call admissions warm leads (cold outreach)
Reason: Brand standard — inbound-led only, no cold contact.
Decision logged: Nova restricted to re-engagement of past inquiries only.
Intel settings · click any to edit · changes audit-logged