1. Separate the basket selloff from the business: check segment growth through the scare
The repeatable method
- When a whole category is sold on a narrative (here, "AI kills software"), list the names that fell with it.
- For each, pull the first reported half after the selloff in constant currency, segment by segment, plus a per-unit pricing metric (revenue per booking, per seat, per user).
- Keep only names where every segment grew and unit pricing rose despite volume headwinds — the market sold the label, not the numbers.
Here: AMS.MC fell in January's SaaSpocalypse, then printed H1 revenue +5% cc, adj. EPS +7.3%, Hospitality +9.2%, Air IT +8.7% and revenue per booking +5.1% — even with slower bookings, the Iran war and IATA cutting its traffic forecast.
Watch for
- Unit pricing holding while volumes slow; customer wins announced during the selloff; management buying assets while peers are cheap.
2. Test the AI-disruption claim with the switching-cost question
The repeatable method
- Ask what happens to the customer if the software fails for an hour — is the lost inventory recoverable?
- Ask whether the customer could build it in-house at a sensible price, given regulation and error tolerance.
- If failure is unrecoverable and in-sourcing is uneconomic, AI is more likely a feature the incumbent adds than a replacement.
Here: "Seats not sold today vanish forever" (Harding Loevner); "No airline can build a system with the same capabilities at a sensible price" (GMO); "New entrants struggle in an industry with low tolerance for error" (Morgan Stanley IM). The article's named risk is exactly the opposite case — airlines in-sourcing IT and going direct.
Watch for
- Large customers announcing in-house builds or direct-channel shifts; contract renewals and churn; regulatory changes lowering error costs.
3. Blend two independent valuation methods and check leverage against the nearest peer
The repeatable method
- Run a conservative DCF (state terminal value, WACC, and assume flat FCF growth).
- Run a re-rating case: the multiple the stock held before the scare, applied to a conservative current-year EPS.
- Average the two for the target; then compare net debt/EBITDA with the closest competitor to judge downside if the thesis is wrong.
Here: DCF with flat FCF → $111; last summer's 21x multiple on ~$3.70 EPS → $77; average $94 on the ADR, ~40% upside. Balance sheet ~1x net debt/EBITDA vs SABR at 7.5x, with an 8.4% FCF yield as cushion.
Watch for
- Forward multiple drifting back toward the pre-scare level; FCF yield compressing; any leverage added for M&A (the €1.2B Idemia deal).