TicketSwap

Finding

TicketSwap confirmed in an AWS case study that their EKS cluster runs on a baseline of approximately 10 nodes in autumn and winter, scales 10x to roughly 100 nodes during peak event season, and hit 15x capacity instantaneously when a major event sold out. The platform has run on AWS since approximately 2013-2014: over a decade. No infrastructure cost review or optimization program has ever been disclosed publicly. Finance is outsourced to Koppel Services with no indication of an internal FinOps function. Total capital raised across 13 years is $12.2M, a lean base that does not suggest capacity for infrastructure optimization work alongside rapid international expansion.

Ten nodes in autumn. A hundred and fifty at instantaneous peak. A deployment that has run for over a decade with no disclosed cost review.

Our Solution

Lightbloom runs a full AWS cost audit using Cost Explorer and Trusted Advisor: right-size baseline EKS node groups against actual utilisation data, identify Savings Plan or Reserved Instance coverage for steady-state compute, tag and allocate unattributed resources, and model the optimal Spot strategy for peak-season burst. The deliverable is a cost-attributed AWS deployment with automated monthly spend reporting in Looker, confirmed with the engineering team before any infrastructure changes are made. For steady-state baseline compute, 1-year Savings Plans typically reduce on-demand costs by 30-40%.

Finding

TicketSwap's support operation runs on Intercom and Aircall, confirmed via open-source Meltano taps on GitHub. Trustpilot and PissedConsumer reviews document a consistent pattern: disputed transactions take 2–3 weeks to resolve, handled entirely through web forms and chat with no direct phone or email contact available. For a platform with 18 million registered users and millions of annual transactions across 42 countries, the dispute queue drives a significant share of support agent time. Most TicketSwap disputes are binary: SecureSwap's barcode audit log either confirms the new ticket was scanned by the venue (dispute invalid) or confirms it was never scanned (refund warranted). The 2–3 week resolution window is not driven by legal complexity. It is the time a human agent takes to query Stripe Connect payout status, the SecureSwap transfer log, and event-side barcode records, then draft a response in the buyer's language.

The barcode either scanned at the venue or it did not. The log knows. A human still waits three weeks to check it.

Our Solution

Lightbloom builds a dispute triage agent integrated with Intercom, the Stripe Connect API, and TicketSwap's SecureSwap transfer audit log. The agent classifies each incoming dispute, queries the relevant data sources, and generates a decision and response draft for clear-cut cases. Ambiguous cases route to a human agent with a pre-completed evidence summary. Expected outcome: 50–60% of disputes resolved end-to-end automatically, reducing mean resolution time from 2–3 weeks to under 4 hours for automated cases.

Finding

TicketSwap's organizer partner network exceeded 6,000 partners across 42 countries as of 2024, growing at 33% year-on-year. Each partner goes through a structured onboarding flow: account setup, API or widget integration, FairShare pricing configuration, and payout account verification via Stripe Connect. FairShare ties each organizer's buyer fee to their historical sell-through rate, which means reconciliation is not static: it recalculates per event and per transaction cycle. For an international network operating across 42 countries and multiple currencies, that reconciliation load grows linearly with the partner base unless automated. No partner success or operations role is named in any public source. The 2023–2024 headcount contraction reduced capacity precisely as partner growth was accelerating.

Two thousand new organizers a year. Each one needs intake, technical setup, and recurring reconciliation. None of that is automated.

Our Solution

Lightbloom builds an organizer partner operations agent that handles three distinct workflows: structured onboarding intake (account validation, API key provisioning, widget configuration, Stripe Connect payout verification), integration support triage (classifying inbound partner queries, resolving known issues automatically, escalating custom requests to a human), and FairShare reconciliation reporting (per-partner, per-event sell-through calculation, automated fee schedule updates, and monthly reconciliation summaries pushed to each organizer). The result is a partner ops function that scales with the network without scaling headcount.