Skip to content

What-if scenarios

The Scenarios page (/scenarios/, Team and above) answers one question: if conditions change, which of your automations still earn their keep? It re-evaluates your as-reported ledger under conditions you choose (an AI repricing, a haircut on your assumptions, a drop in volume) and shows the result per automation and across the portfolio.

It is read-only and hypothetical. Nothing you do on this page changes a booked number; closed months stay frozen. The page carries the standing label "Hypothetical. Your ledger is unchanged." for exactly that reason.

The one idea to hold onto

Every scenario is a set of multipliers on the numbers already in your ledger. The baseline column reproduces the ledger for the window, to the cent, so you are stressing the number finance already accepted, not inventing a new estimate. When a lever is at ×1 (its default), the scenario equals the ledger.

The three levers

Set them in the controls at the top of the page and press Recompute (or click a preset). They combine: the bear-case preset moves all three at once.

Lever The question it answers What it scales
AI price ×m "My provider reprices ×3 (or moves my flat-rate plan to usage-based): which agents still pencil out?" the metered AI cost of each run
Baseline ×h "My minutes-and-rate assumptions were 30% too generous: does the portfolio survive its own skeptic?" the gross labor value (the assumed side)
Volume ×v "Adoption stalls 30% (or doubles): which automations only work at today's volume?" volume-linked value and run costs (fixed costs do not scale)

A few notes on what each lever does not touch, so the numbers stay honest:

  • AI price scales only the metered AI slice of run cost. Non-AI run costs and fixed costs are unaffected.
  • Baseline applies to all labor value, because minutes × rate is an assumption in every lane. It stacks on top of the conservatism factor already booked on the productivity lane; it does not replace it.
  • Volume scales the per-run economics (value and run cost) proportionally. Fixed costs (build, recurring licenses, one-time fees) are held constant, which is what makes a volume floor appear.

The flat-rate toggle

Agents that run on a flat-rate plan (Claude Max seats, Copilot) book $0 marginal cost in your ledger (that is what you actually pay per run). But LumaTrack also records each covered call's API-equivalent cost as the repricing exposure. Tick "Bill flat-rate usage at list" and the scenario prices that exposure at the AI multiplier, so you can see what happens if the flat-rate era ends for you and that usage moves to metered billing. Leave it off to see today's reality (flat-rate usage at $0).

Presets

The chips across the top are shortcuts:

  • Baseline: everything at ×1 (equals your ledger).
  • AI ×3 / AI ×5: an AI repricing of that multiple.
  • Assumptions −30%: baseline haircut to ×0.7.
  • Volume −30%: volume to ×0.7.
  • Bear case: AI ×3, assumptions ×0.7, volume ×0.7, and flat-rate billed at list, all together. The pessimistic corner.

You can also type any multiplier into the lever boxes (0.1 to 1000) and choose a window of the last 3, 6, or 12 months.

Reading the results

The headline cards show the portfolio's net value as reported (your ledger today), its net under the scenario, the delta between them, and how much AI exposure sits in the window (metered vs riding flat-rate plans).

The per-automation table is sorted most-affected first: the automations the scenario hits hardest are on top. Each row shows its baseline net, its scenario net, and up to three survival badges:

  • AI survival: the repricing multiple the automation survives, e.g. "survives AI ×7.4". Read it as: it stays net-positive until AI prices reach that multiple. "dies at ×1.3" means it is already under water at the scenario's current AI multiplier. "survives any plausible AI repricing (×100+)" means AI cost is a rounding error for it. "not AI-exposed" means it uses no metered AI. The multiple is rounded down, so it never overstates survival.
  • Assumption margin: "assumptions can be 27% too generous" means the automation stays net-positive even if your minutes/rate baseline was overstated by that much. "no assumption margin" means it is already at the edge.
  • Volume floor: "volume floor 22%" means it needs at least 22% of today's volume to cover its fixed costs. "no volume floor" means it has no fixed cost to cover, so any volume works.

For AI-exposed automations the row also names the model mix (the models whose tokens it runs on), so you can see where the exposure comes from.

How to use it

  • Before a renewal or a budget review: run the bear case and point to the automations that survive it; that is the defensible floor.
  • When a provider announces a price change: set the AI multiplier to the change and read the survival badges. Anything showing "dies at ×N" below the announced multiple needs attention (a cheaper model, a flat-rate plan, or retirement).
  • When someone questions your assumptions: hand them the baseline lever. If the portfolio is still positive at ×0.7, the number survives their skepticism.
  • When you run flat-rate agents: tick the flat-rate toggle to see your exposure if those plans move to usage-based billing (several vendors have already done this).

From an agent (MCP)

The same engine is available to agents through the MCP tool get_scenario_stress (see the MCP server section). An agent can ask, for example, whether it would still be net-positive if its own tokens cost five times as much, and get the portfolio and per-automation nets plus breakevens back as structured data.

Availability

Scenarios are a Team and above feature. On the Free tier the page shows a short explainer instead of the tool.