Competitive strategy exercise

Can you protect this business from AI disruption?

Run an established software company or the AI challenger attacking it. Four quarterly decisions, one commitment each. A deterministic engine resolves customer choice, revenue, delivery cost, investment and cash, and shows you the arithmetic behind every number.

Start with a briefing Inspect the model

No account, no sign-up, no email. One click to the role choice.

How a quarter works

Read the public intelligence for the quarter.
Pick one commitment and preview its cost, capacity, timing and risk.
Record an optional rationale.
Lock the decision. Only then does the opponent's move appear.
Read the resolution, line by line, with links to the rules.

Afterwards you can rewind any decision into a branch, or switch sides and play against the strategy you just recorded.

The two sides

Incumbent

RelayWorks

Project-management seats, sold per team per year.

Opening cash: $3,000,000
Annual price, small / enterprise team: $2,400 / $24,000
Teams today: 400 small, 120 enterprise
Delivery capacity: 3 effort units per quarter

Challenger

TaskPilot

Completed work bundles, quoted as a fixed annual amount at the segment's assumed usage.

Opening cash: $1,000,000
Annual price, small / enterprise team: $1,800 / $18,000
Teams today: 80 small, 5 enterprise
Delivery capacity: 2 effort units per quarter

What the units mean

Contribution: Revenue less the variable cost of delivering to those customers. It does not include fixed operating cost.
Customer retention: The share of a supplier's customers who do not reconsider in a quarter, plus those who reconsider and choose the same supplier again.
Cash: Opening cash minus investment outlays plus operating cash flow. This exercise reports operating cash flow, not EBITDA and not audited profit, because the model is too simple to deserve either label.
Quarter: One decision period. Prices are annual per team, so quarterly revenue is the annual price divided by four.

Evidence status

This is a concept prototype. The playable product, its documented rules and its exported runs are the evidence. There are no playtest findings, usage statistics, or user counts to report, because no playtests have been run. The analysis page contains an explicitly labelled illustrative draft rather than observed results.