P
PalmStat ERP · Palm Stationery
Built by TW Technology Solutions (Pty) Ltd
PROPOSAL · CONFIDENTIAL

PalmStat ERP · v2 Roadmap

From digitisation to intelligence · the next 18 months. A four-phase plan to evolve PalmStat from a system of record into a system of insight.
Prepared for
Palm Stationery Mfrs (Pty) Ltd
Prepared by
TW Technology Solutions
Document date
09 May 2026
Validity
14 days
Where we are today

v1 has digitised the lifecycle. v2 makes it think.

7
Lifecycle steps live
100%
Documents captured
36.8m
RM value tracked (R)
18
Months · v2 horizon
2
Power Platform
Months 1–3
3
AI & Predictive
Months 4–9
4
Portals
Months 10–12
5
IoT & Twin
Months 13–18
Executive summary. v1 of PalmStat ERP digitised the seven-step production lifecycle from GRN to delivery, captured every supporting document, enforced data integrity through reference data and SOPs, and gave Palm Stationery a single source of truth for raw material, WIP, and finished-goods movement. The foundation is now solid — tally sheets, machine breakdowns, paper movement, and dispatch all flow through one system. v2 layers intelligence on top of that foundation: Power BI dashboards for executives, Power Automate flows that remove manual hand-offs, AI models that predict failures and yields before they happen, customer and supplier portals that drive retention, and IoT telemetry that closes the loop between the floor and the boardroom. The combined investment of R 1.73m once-off over 18 months is recovered inside 14 months through measurable efficiency, yield, downtime, and retention gains.
PalmStat ERP · v2 RoadmapPage 1 of 9
Phase 2 of 4

Power Platform integration

Three months to put a Microsoft-native intelligence layer over PalmStat — dashboards executives actually open, automation that removes manual email chains, and mobile companions that take the system off the desk and onto the shop floor.

Phase 2 · Power Platform

Dashboards, automation, mobile

Move from "the data is in PalmStat" to "the data finds the right person at the right time" — using the M365 stack Palm already pays for.
Months 1–3 3 months R 280k once-off
Power BI · 5 dashboards
  • Production efficiency — daily output vs plan, by machine and shift
  • OEE board — availability × performance × quality, drilled to operator
  • Yield by SKU — RM consumed vs FG produced, variance flagged in red
  • Dispatch performance — OTIF, lead time, customer-level breakdown
  • Customer profitability — revenue, margin, returns, days-to-pay
Power Automate · 12 flows
  • Tally sign-off → auto-email to ops manager
  • GRN approval → sync to Xero and Sage
  • Low-stock RM → Teams alert + draft PO
  • Shift schedule → M365 calendar invites for operators
  • Breakdown logged → SMS to maintenance lead
  • + 7 more flows across QC, dispatch, and HR
Power Apps · 3 mobile companions
  • Shop-floor app — tally entry, breakdown logging, photo capture
  • Executive briefing app — daily KPI cards, push notifications
  • Supplier portal app — PO confirmation, delivery slot booking
  • SSO via Microsoft Entra — no extra logins
InvestmentR 280,000 once-off
Run costR 8,500 / month per app · M365 P1/P2 licensing
Outcome1,200 manual hours / yr removed · instant exec visibility
OwnerIveshan Naidoo · TW Technology Solutions

All three pillars share the same Dataverse-mirrored model that PalmStat already exposes. No re-keying, no swivel-chair integration, no third-party connectors with month-to-month risk — everything Palm pays for, used the way it was meant to be used.

Phase 2 · Power Platform integrationPage 2 of 9
Phase 3 of 4

AI & predictive intelligence

Six months to take three years of PalmStat history — tally sheets, breakdowns, sales, paper movement — and turn it into models that look forward instead of back. Not a "future feature": a working forecast on every shift planner's screen.

Phase 3 · AI + ML

Predict before it costs

Four production-grade models, deployed against live PalmStat data on Azure, owned by Palm. No black-box SaaS, no data leaving the tenant.
Months 4–9 6 months R 450k phase setup
Predictive maintenance
  • ML model trained on machine_breakdowns + shift logs
  • Flags machines likely to fail in next 7 days
  • Pushes work orders to workshop_repairs queue
  • Saves est. 12% downtime · ~96 hrs / yr
Demand forecasting
  • Time-series on sales_orders, by SKU and customer
  • Auto-suggested 4-week production plan
  • Triggers raw-material reorder levels dynamically
  • Reduces over-stock and stock-outs simultaneously
Yield optimisation
  • Regression on tally + breakdown + paper grade
  • Suggests optimal machine + shift + operator combo
  • Flags poor-yield combos to ops manager in real-time
  • Target +1.5% yield on R 36.8m RM
Quality vision (Azure AI)
  • Camera feed on QC station — auto-detect print defects, mis-cuts, ink spread
  • Azure AI Vision custom model trained on Palm's reject library
  • Defect score auto-attached to FGRN; consistent reject reporting
  • Reduces QC reliance on individual operator judgement
Phase setupR 450,000 once-off
Azure consumptionR 12,000–18,000 / month · usage-based
OutcomeDown 12% · Up 1.5% yield · defect detection live
Data residencyAzure South Africa North · POPIA-aligned

Models are retrained monthly against the previous 90 days of operational data, with drift monitoring on Azure ML. If a model degrades, the dashboard says so — we never let a stale forecast drive a real decision.

Phase 3 · AI & predictive intelligencePage 3 of 9
Phase 4 of 4

Customer & supplier portals

Three months to put a polished, branded face on PalmStat — one for the customers placing orders, one for the suppliers feeding the floor. Both reduce inbound phone-calls, both drive retention through self-service.

Phase 4 · Portals

Self-service for both sides

Customers track orders, download invoices, raise queries. Suppliers confirm POs, upload certificates, see when they'll be paid. PalmStat becomes the front door, not the back office.
Months 10–12 3 months R 380k once-off
Customer self-service portal
  • Live order status — from sales order to dispatch tracking
  • Self-serve invoice and statement download (PDF + CSV)
  • Raise issues / returns / re-orders without phoning sales
  • Branded under Palm Stationery, hosted on palmstat.co.za
  • Drives retention — estimated +R 1.2m / yr year 2 onward
Supplier portal
  • Live GRN feed — suppliers see what was received and when
  • PO confirmation flow with delivery slot booking
  • Payment status visibility — reduces "where's my money?" calls
  • Certificate-of-conformance upload, attached to GRN automatically
Identity & security
  • SSO via Microsoft Entra B2B — no shared passwords
  • Per-customer / per-supplier scoped views — no data bleed
  • Multi-factor authentication mandated for every external user
  • Full audit log on every download, every action
  • POPIA-compliant by design
Build costR 380,000 once-off
Hostingincluded in existing M365 / Azure spend
Outcome Y2++R 1.2m retention · -40% inbound calls
BrandingPalm Stationery · not TW · not generic

Both portals are PWA-style: they install to the home screen on a phone, work offline-tolerant, and look native on every platform. No App Store review cycles, no Android-vs-iOS divergence.

Phase 4 · Customer & supplier portalsPage 4 of 9
Phase 5 of 4 · Stretch

IoT, telemetry & the digital twin

Six months to put sensors on the top five machines, stream their state to PalmStat in real-time, and bring up a 3D digital twin of the Pstat plant that lets you click any machine and see live KPIs. The visible payoff for the boardroom — and the foundation for energy optimisation.

Phase 5 · IoT + digital twin

Close the loop between floor and dashboard

Every dashboard so far is post-hoc. With IoT, KPIs update as the machines breathe — and a 3D twin makes that update visible at a glance, even from the boardroom screen on the second floor.
Months 13–18 6 months R 620k hardware + setup
IoT sensors
  • Vibration, temperature, current draw on top 5 machines
  • Edge gateway per line — no single point of failure
  • Azure IoT Hub ingestion · ~1,200 events / minute
  • Real-time machine health score on Live Floor view
Digital twin
  • 3D model of the Pstat plant, accurate to layout
  • Click any machine → live KPIs, current job, current operator
  • WebGL — works in any browser, on any screen
  • Boardroom-grade visualisation for monthly reviews
Energy optimisation
  • Track kWh per pad, per shift, per machine
  • Suggest off-peak runs to dodge Eskom peak tariffs
  • Carbon-per-pad reported alongside cost-per-pad
  • Foundation for future ESG reporting and B-BBEE submissions
Hardware + setupR 620,000 once-off
Telemetry runR 4,000 / month · Azure IoT Hub
OutcomeReal-time floor · energy savings · ESG-ready
Future-fitFoundation for Phase 6 · full predictive plant

The twin doubles as a sales tool: a 60-second walk-through of the live plant, on a tablet, in front of any prospective wholesale customer. Few SA stationery manufacturers have anything close.

Phase 5 · IoT & digital twinPage 5 of 9
Investment summary

Total v2 investment

Combined cost of all four phases over 18 months. Once-off costs are billed at phase kick-off. Monthly costs are the run-rate after the phase is live; they accrete as each phase ships.

Phase Theme Months Once-off (R) Monthly run (R) Year 1 total (R)
Phase 2 Power Platform · BI, Automate, Apps 1–3 280,000 8,500 357,000
Phase 3 AI / ML · predict, forecast, yield, vision 4–9 450,000 15,000 585,000
Phase 4 Customer & supplier portals 10–12 380,000 0 380,000
Phase 5 IoT, telemetry, digital twin 13–18 620,000 4,000 644,000
Grand total v2 · over 18 months 1,730,000 27,500 avg ~ 2,025,000
R 1.73m
Once-off · full v2
~R 27.5k
Monthly run · avg blended
~R 2.03m
Year 1 total cost
14 mo
Payback period

All figures exclude VAT. Once-off costs are billed monthly across each phase's runway, against agreed milestones. Monthly costs invoice on the 1st in arrears. No phase commits Palm to the next — each is signed off independently.

Phasing protects cash flow

Phase 2 alone (R 280k) unlocks 60% of the Year-1 efficiency gains. Palm can stop after any phase, keep the value already shipped, and revisit later.

Microsoft licensing leverage

Palm's existing M365 Business Premium seats include Power Apps, Power BI Pro, and Automate at no extra cost for internal users. Run-cost above is incremental only.

Investment · combinedPage 6 of 9
Return on investment

Payback in 14 months

All gain figures are conservative — based on the lower bound of measurable improvements seen at peer SA manufacturers running comparable Power Platform + AI rollouts. We don't take credit for gains we can't measure inside PalmStat itself.

Cumulative ROI · year 1, 2, 3 (R '000)

3,200 2,400 1,600 800 0 Year 1 Cost 2,025 Gain 1,061 Year 2 Cost 330 Gain 2,261 Year 3 Cost 330 Gain 2,710 Investment Annual benefit Payback · month 14

Year 1 benefit breakdown (R)

Efficiency · 1,200 hrs × R 280 R 336,000 Yield · +1.5% on R 36.8m R 552,000 Downtime · 12% × 800 hrs × R 1,800 R 173,000 Retention · portal · from Y2 + R 1.2m / yr Y2+ Year 1 total benefit R 1,061,000

Year 1 measurable gains

  • Efficiency — 1,200 hrs / yr removed @ R 280/hrR 336,000
  • Yield — +1.5% on R 36.8m RM throughputR 552,000
  • Downtime — 12% × 800 hrs × R 1,800/hrR 173,000
  • Year 1 totalR 1,061,000

Year 2–3 compounding

  • Customer portal retention+R 1.2m / yr
  • Forecast accuracy · reduced over-stock+R 380k / yr
  • Energy · off-peak shift mix+R 290k / yr
  • Year 2 run-rateR 2.93m / yr
Payback period
14 months
Y3 cumulative
R 4.0m+
ROI · payback & benefit detailPage 7 of 9
Risks & dependencies

What could slow this down

We'd rather list the risks up-front than pretend they don't exist. Every item below has a named mitigation in the Master Services Agreement.

Power Platform licensing

Power Apps per-app plans are R 8.5k / month each. Mitigation: M365 Business Premium seats already include core Power Apps for internal users; we re-use those before adding paid plans.

Azure consumption

AI / IoT phases run on metered Azure services. Mitigation: monthly cap of R 18k on AI workloads + R 4k on IoT, alerted via Cost Management; auto-throttle on breach.

IoT hardware lead time

Industrial-grade vibration + current sensors carry 8–12 week lead time from EU. Mitigation: orders placed at month 11 of Phase 4, before Phase 5 kick-off.

Operator training

New mobile apps and floor changes need adoption. Mitigation: 2-day on-site rollout per phase, video SOPs added to PalmStat SOP Library, super-user nominated per shift.

Data quality during AI training

Models inherit any gaps in tally / breakdown logging. Mitigation: 60-day data audit before Phase 3 kick-off; gaps closed via Phase 2 Power Automate prompts before training begins.

Change control

Each phase introduces new screens. Mitigation: feature-flag rollout, pilot with one shift first, full rollout only on operator sign-off; rollback capability on every release.


Decision matrix

Pick your tier

Four ways to engage v2. Each tier delivers value on its own; later tiers can be added at any point without re-buying earlier work.

Tier 1 · entry
Bronze
R 280k once-off
  • Phase 2 · Power Platform
  • Phase 3 · AI / ML
  • Phase 4 · Portals
  • Phase 5 · IoT / twin
  • Year 1 gain: R 509k
Tier 2
Silver
R 730k once-off
  • Phase 2 · Power Platform
  • Phase 3 · AI / ML
  • Phase 4 · Portals
  • Phase 5 · IoT / twin
  • Year 1 gain: R 1.06m
Tier 3
Gold
R 1.11m once-off
  • Phase 2 · Power Platform
  • Phase 3 · AI / ML
  • Phase 4 · Portals
  • Phase 5 · IoT / twin
  • Y1 gain: R 1.06m + Y2 +R 1.2m
Recommended
Tier 4 · full v2
Platinum
R 1.73m once-off
  • Phase 2 · Power Platform
  • Phase 3 · AI / ML
  • Phase 4 · Portals
  • Phase 5 · IoT / twin
  • Y3 cumulative: R 4.0m+
Risk & decision matrixPage 8 of 9
Approval & sign-off

Authorisation to proceed

Signature against the chosen tier authorises TW Technology Solutions to issue a Statement of Work and begin Phase 2 within 14 calendar days. Tier selection can be amended in writing prior to first invoice.

Item Detail
Selected tier□ Bronze  ·  □ Silver  ·  □ Gold  ·  □ Platinum (recommended)
Once-off investmentAs per tier above · ex VAT · billed monthly against milestones
Monthly run costFrom R 8,500 to R 27,500 / month, accreting per phase live
Phase 2 startWithin 14 calendar days of countersignature
Governing agreementTW MSA 2025-A · with PalmStat-specific Schedule B (data protection, POPIA)
ValidityThis proposal is valid for 14 days from document date
For Palm Stationery
Palm Stationery Board
CEO / FD authorised signatory
Signature
Date
Solution Lead
Iveshan Naidoo
Head of Delivery · TW Technology Solutions
Signature
Date
For TW Technology Solutions
Brent Reddy
Chief Executive Officer · TW Group
Signature
Date
Next step. On countersignature, TW will issue the Phase 2 Statement of Work within 5 working days. Phase 2 kick-off workshop is scheduled on the same day, on-site at the Pstat plant, with Iveshan, the ops manager, and the IT lead. The first Power BI dashboard ships in week 4. The first Power Automate flow goes live in week 6. We don't ask Palm to wait three months to see the first piece of v2 in production.
Generated 09 May 2026 · Built by TW Technology Solutions (Pty) Ltd · Confidential proposal · Document ID: PSTAT-V2-RM-2026-05
Sign-off & next stepsPage 9 of 9