Living artifact Owner: Marcus Higgins· Audience: ELT· Target: Sep 30, 2026

Eliminating $50K in SaaS spend by September 30, 2026

Landing caught up to the speed of AI in how we build, serve, and operate, but we have not yet harvested the cost savings. Engineering will always prioritize revenue and guest experience over cutting SaaS, so the objective is for business leaders to rebuild their own small tools. The honest read on the data: the big dollars are moats you cannot easily rebuild; the easy wins are the small line items. The number below is real, not inflated.

Committed annual savings
$0
Goal
$50,000
Gap to goal
$50,000
0% of goal$50,000
60%
Spend confirmed against Ramp: 615 software transactions, Jan 1 to Jun 15, 2026. Renewal column is payment cadence from Ramp; Ramp holds no contract-agreement records, so specific dates are unavailable and the 2024 vendor sheet is stale. Most candidates are month-to-month, so they can be cut anytime.
Tier 1 — high confidence, easy, low risk
Tier 2 — buildable, partial capture (scaled by the slider)
What we will not touch (about 80% of spend)

Moats, infrastructure, or regulated and security-sensitive systems. Rebuilding these with Claude is not easy and would cost more than it saves: PriceLabs ($84,918, dynamic pricing engine), Rentals United ($59,041, OTA channel manager), Time Doctor ($24,048, workforce monitoring), Koddi ($20,000, ad bidding), cloud and dev infra (about $30k+), and the sticky systems of record (Salesforce, HubSpot, Slack, Dashlane).

Why this is a Claude Artifact, not a memo

What I care about most is using Claude Artifacts for future ELT meetings. A frozen forecast gets approved or rejected as a single point. A live model gets pressure-tested across ranges in minutes, which is how we actually think. Drag the capture slider above and watch the gap move: that is the whole argument made tangible.

Less in-room load. The artifact answers the clarifying questions so the sponsor stops being a human query engine and the hour goes to the real debate.
Better quantitative decisions. Ranges over absolutes, downside cases explicit. You catch fragility (a plan hinging on clean ramp) faster.
Data integrity becomes structural. Numbers cite a source or get flagged as estimates, enforced by the tool instead of by trust.

How to create a Claude Artifact

Data integrity: dollar figures are annualized from confirmed Ramp card transactions (615 records, Jan 1 to Jun 15, 2026). Ramp holds no contract-agreement records, so renewal cadence is inferred from payment frequency and any specific dates remain unverified. Treat the annual totals as current run-rate estimates, not committed contract values. Built using Claude Artifacts.