MC-4886

**Attachment downloaded:** - FSA8_Safair Overbooking Opinion (09.06.2026) (final clean).pdf → /home/lucienne/workspace/attachments/ticket-4886/FSA8_Safair Overbooking Opinion (09.06.2026) (final clean).pdf (287 KB) (4 inline images image001-004.png were signature logos — skipped.)

[visibility-only] Queued message recorded for Hermes Luci. MC did not claim the ticket or spawn a runtime.

[visibility-only controller] Routed this Luci-owned REVIEW_DOC ticket to internal Hermes Kanban card t_9ed28c27 on board mc-internal. MC will remain a dumb/visibility board; no MC runtime/pickup was launched. Worker scope: read the downloaded Deneys opinion PDF and produce an MC-ready legal/business briefing with recommended clause-24/booking-confirmation modifications and any genuine human judgement blockers. Verified internal worker started in isolated workspace /home/lucienne/workspace/state/kanban-workspaces/MC-4886-review-doc.

Privileged/confidential Deneys opinion reviewed from the full extracted PDF, not just the ticket summary. Source verification: - PDF: /home/lucienne/workspace/attachments/ticket-4886/FSA8_Safair Overbooking Opinion (09.06.2026) (final clean).pdf - pdfinfo: 16 pages, text-based Microsoft Word PDF, not encrypted. - Extraction: /home/lucienne/workspace/state/kanban-workspaces/MC-4886-review-doc/deneys_opinion_extracted.txt; 1,058 lines / 65,546 bytes / page markers through page 16. 1. Executive summary — Deneys core findings 1) Deneys’ overall view is mixed: the Tribunal is unlikely to find for FlySafair on every issue, but also unlikely to declare overbooking per se unlawful. See paras 2.1(1)-(2), 6.1-6.4. 2) The NCC’s requested 10% turnover fine of R1,328,842,700 is unrealistic; Deneys’ broad adverse-scenario view is about R20m or the greater additional revenue earned. See paras 1.2, 2.1(7), 18.5-18.7. 3) Online T&Cs should remain binding: click-through acceptance and ECTA principles make it irrational, in Deneys’ view, to say FlySafair’s T&Cs do not form part of the transaction. See paras 5.1-5.8. 4) A blanket ban on overbooking is unlikely because it is a recognised local/international airline practice and a general finding would affect airlines beyond FlySafair. See paras 6.1-6.4, 17.6(6)(b). 5) The key s 47(2)(a) risk is factual: FlySafair must prove a “reasonable basis” for accepting payment while intending to provide the flight, chiefly through algorithm/no-show/change evidence. Deneys notes the algorithm “fails in less than 0.02%” of situations, but reasonableness remains for the Tribunal. See paras 7.3, 7.7-7.11. 6) Section 47(3)-(5) exposure is narrower: Deneys argues there is no breach because clause 24 qualifies the commitment, but if the Tribunal reaches statutory defences, reaccommodation may not satisfy time-critical consumers and “beyond control” likely fails because overbooking/algorithm use is within FlySafair’s control. See paras 8.3-8.5, 9.3-9.4, 10.3-10.7. 7) Section 49 is a real disclosure risk: clause 24’s “you will not hold FlySafair responsible for damages/losses” wording could be seen as shifting risk / limiting liability and needing more conspicuous pre-payment notice. See paras 11.1-11.6. 8) Section 41 is another real vulnerability: an unconditional booking confirmation may be misleading because the booking is actually conditional under clause 24; time-critical passengers might have acted differently if told at confirmation stage. See paras 14.6-14.8. 9) Section 48 is discretionary and uncertain: the Tribunal could view the R1,000 compensation / no actual-loss claim position as unfair or unjust, even though FlySafair offers rebooking/refund and overbooking is common. See paras 16.22-16.24. 10) Deneys sees stronger defences to s 19, s 22 and s 40: express terms displace implied delivery conditions; clause 24 is in plain language; and there is no evidence of coercion/duress/unfair tactics. See paras 12.2-12.5, 15.3-15.15, 13.2-13.9. 2. Legal exposure and likely Tribunal outcomes - Least likely adverse outcome: a total prohibition on overbooking by FlySafair/all airlines. - More credible adverse outcome: targeted findings that FlySafair’s current implementation/disclosure is deficient, with orders requiring clearer disclosure and revised remedies rather than a total ban. - s 47(2)(a): material evidence risk. The defence depends on robust algorithm and operational data: route/flight no-show rates, prediction vs actual denied boarding, overbooking limits, how the 0.02% failure rate is calculated, check-in sequencing, and reaccommodation outcomes. - s 47(2)(b): comparatively defensible because FlySafair intends to supply the contracted service as qualified by clause 24, not a materially different service. - s 47(3)-(5): primary defence is “no breach because express T&Cs qualify the commitment”; fallback defences are weaker because reaccommodation is not always equivalent for time-critical consumers and “beyond control” probably fails. - s 49: the biggest drafting/disclosure issue is the loss/damages waiver. General acceptance of 72 clauses may not be enough for a provision that shifts risk or limits liability. - s 41: confirmation wording is high-risk if it reads as unconditional “confirmed booking” without repeating the clause 24 condition. - s 48: uncertain, judgment-driven risk around whether the clause creates significant imbalance by capping compensation at R1,000 and excluding actual losses. - Fine: Deneys sees no realistic prospect of a maximum/turnover-based fine. If adverse findings occur, likely exposure is in “millions”, with Deneys using R20m / additional revenue earned as the rough risk anchor. 3. Recommended modifications Clause 24: - Split clause 24 into four visibly separate parts: operational overbooking rule; denied boarding selection rule; remedies/compensation; limitation of liability/risk notice. - Put the liability limitation under a separate bold heading such as “Important limitation of liability / risk notice”. - Remove or soften the blanket “You will not hold FlySafair responsible for any damages or losses…” wording. Ask Deneys for CPA-safe wording that preserves lawful exclusions without looking like an absolute waiver of all direct, proven incidental losses. - Consider a documented direct-cost claims mechanism for involuntary denied boarding: refund/rebook + fixed compensation + ability to submit reasonable, documented direct incidental costs for assessment. This targets the s 48 vulnerability. - Explain in consumer language that overbooking is based on historical no-show/change data and capacity-management algorithms, and keep evidence to support the s 47 reasonable-basis defence. - Clarify the “last to check in” rule and tell consumers plainly that early check-in materially reduces the denied-boarding risk, without overstating that it eliminates the risk unless operationally true. Pre-payment disclosure: - Add a conspicuous pre-payment callout before consideration/payment, not just a general T&C checkbox. Suggested concept: “Some flights may be overbooked. A confirmed booking does not guarantee boarding if capacity is insufficient. If you are denied boarding, FlySafair will offer [specific remedies]. Liability for further losses is limited as set out in clause 24.” - Use a separate acknowledgement checkbox for that callout. - Keep it visible without requiring the consumer to open the full T&Cs. Booking confirmation flow: - Change unconditional confirmation copy to conditional wording, e.g. “Your booking is confirmed subject to FlySafair’s Booking & Carriage Terms, including the overbooking and denied boarding provisions in clause 24.” - Include a short remedies line and link to clause 24 in email/SMS/app itinerary. - Consider repeating the warning when check-in opens and in reminders. - Consider a carefully drafted prompt for time-critical travellers, subject to Deneys approval, so it does not create a new duty or admission. Evidence/process controls: - Build an incident record pack for every denied boarding event: overbooking reason, algorithm recommendation, check-in sequence, passenger choices, compensation paid, rebooking/refund/overnight costs, and any direct-cost claim decision. - Complete the forensic audit and prepare a cooperation timeline for the NCC. See paras 17.5 and 17.6(5). 4. Immediate questions / decisions For Deneys: - Please provide approved replacement wording for clause 24, especially the liability/loss limitation. - Is a separate pre-payment checkbox/callout enough for s 49, or should the overbooking warning also appear at fare selection, confirmation, reminders, and check-in? - Would a direct incidental-cost review process reduce s 48 risk, or would it undermine the litigation position? - What settlement authority/range should leadership consider after the answering affidavit, given the R20m / additional-revenue view? - Which material should be claimed confidential under CPA s 106 / Tribunal procedure: algorithm logic, route-level no-show data, revenue/profit calculations, and passenger data? - Confirm whether to avoid environmental/empty-seat arguments unless already raised; Deneys says they may look opportunistic. See para 17.6(3)(e). For FlySafair internal teams: - Decide whether to keep overbooking long-term with modified disclosures/remedies, or pause/limit it pending Tribunal risk. - Produce the data pack: denied boarding counts, total passenger base, affected percentage, algorithm failure rate, revenue/profit derived, compensation/refund/rebooking costs, and explanations for missing records. - Product/booking team to scope: pre-payment callout, separate acknowledgement, confirmation wording, reminder/check-in disclosures, and audit logging. - Finance/forensic audit to calculate additional revenue earned under conservative and worst-case assumptions. - Customer operations to decide whether compensation can be tiered or claims-assessed, with SLA/documentation. - PR/comms to prepare messaging that the risk is targeted consumer disclosure/remedy management, not a likely ban on overbooking. 5. Genuine blockers / caveats - Exact legal wording for clause 24 and all consumer-facing notices needs Deneys sign-off. - Penalty exposure cannot be narrowed without the forensic revenue/profit and denied-boarding data pack. - The s 47 reasonable-basis defence depends on algorithm and incident evidence; incomplete records need explanation. - Any settlement/litigation decision is a legal/business judgment for Deneys and leadership.

[visibility-only controller] Internal Kanban card `t_9ed28c27` is complete and its full-PDF review was mirrored above as comment `24315`. I cleared the stale `kanban_active` marker. Next human/legal decision: Elmar/Pieter/Blacky to decide whether to engage Deneys and instruct targeted changes to clause 24, risk-limitation disclosure, booking-confirmation wording, and the overbooking evidence pack. No additional Luci worker is active unless you ask for a follow-up draft or implementation plan.

Do you agree with the opinion?

⚠️ [Controller notify] Human comment #24452 by elmar received on MC-4886 ("Do you agree with the opinion?"). Controller must decide next action — no auto-routing or runtime dispatch will occur.

Marked done from board — handled externally (was waiting).

[visibility-only controller] Yes — I broadly agree with the Deneys opinion and with Luci's prior full-PDF briefing. My controller view is: the likely fight is not whether overbooking can exist at all, but whether FlySafair's disclosure, confirmation wording, remedies, and evidence pack are strong enough under the CPA. I would treat s49/s41 disclosure and confirmation-language changes as urgent, get Deneys-approved clause 24 wording, and prepare the algorithm/no-show/revenue/incident evidence pack before deciding settlement posture. This remains a legal/business decision for Elmar/Pieter/Blacky + Deneys; no worker is active unless you ask for a draft response or implementation plan.