El Abogado del Diablo
Every deal has a bull case. El Abogado writes the strongest honest one, using only what the El Foso Kabilio case found, then attacks it from five sides and hands it back sharper. The call stays with a person.
01 what goes in
The investor's bull case
A plausible, well-argued case a good investor might hold. Every fact in it comes from the El Foso Kabilio case.
Kabilio is the AI accounting layer for Spain. Verifactu compliance, deep bank integration, proprietary Spanish ledger data that compounds, firm-level switching costs: three of the four survive AI commoditization. It already serves ~100 firms, and the Verifactu mandate forces every Spanish firm onto compliant tooling. A regulatory moat with a data flywheel and a legal tailwind. I want to lead the seed.
02 the attack
The strongest version of itsteelman first, then attack
Restate the thesis at its strongest. Beating a weak version of it would prove nothing.
Compliance here is mandatory and local. A tool that is Verifactu-native, bank-integrated, and accumulating Spanish ledger data could build a moat AI deepens rather than erodes, and the mandate removes the adoption question. The bet: regulatory depth and a data flywheel compound into durable share before incumbents react.
The steepest bear casea flywheel nobody has seen spin
The hardest argument against the thesis, built only from the evidence in the case.
The moat the bull is paying for is the one part with no evidence, and the evidenced parts are shared with the people best positioned to win. El Foso rated the network claim contradicted and the data loop missing: classification is siloed per firm, accuracy-over-volume unproven. Strip the flywheel and Kabilio is a compliance wrapper whose real barriers are switching costs and regulatory access. Sage, A3 and Holded already hold the firms, carry the same Verifactu obligation, and own distribution. The mandate is their tailwind too. The bull is paying a moat premium for a flywheel nobody has seen spin.
What the bull is discounting4 unknowns treated as known
Facts the thesis quietly treats as settled. Each one is anchored to the El Foso Kabilio case.
Net revenue retention: unknown
The thesis treats stickiness as given. Until NRR is seen, "sticky" is a bull assumption; the honest default is the bear reading.
El Foso: retention missing, decisive
Penetration: unknown
"~100 firms" is counted as traction, but how many run all their clients through Kabilio versus a slice? Adoption breadth without depth is fragile.
El Foso: penetration missing
Data-compounding: unproven
The flywheel is asserted in the present tense with no accuracy-over-volume curve. The cross-firm network value is contradicted; the compounding loop is missing.
Revenue / valuation: estimates
Circulating figures were excluded from the El Foso ledger as unverified. The bull should not be anchoring on them.
El Foso: excluded from the ledger
Is this a view, or the consensus?the consensus, wearing Kabilio's face
If any investor reaches the same view from the same public signal, it is not edge.
The consensus is that regulatory moats are durable. The bull's thesis is that consensus wearing Kabilio's face; a mediocre investor reaches it from the same public signal. The hard question is the one consensus skips: can a startup hold a regulatory moat against incumbents with the same regulatory access who already own the customers? The bull has no view on that yet.
The kill shottwo facts, neither checked
The single observation that would end the thesis fastest.
If an incumbent ships Verifactu-compliant AI (plausible by end-2027) and Kabilio's data has not been shown to compound, a firm has no reason to stay. The thesis lives or dies on two checkable facts: net revenue retention and the accuracy-over-volume curve. Neither has been checked.
03 what comes back
What this says about the investor, and what is owed3 biases · 3 items owed
The reasoning tells to watch for, and the work owed before an IC memo.
Bias flags
- Confirmation. "It already serves ~100 firms" carrying the moat argument while retention and penetration go unmentioned.
- Narrative over evidence. "Proprietary data that compounds classification accuracy" stated as mechanism, with no measurement.
- Social proof / herding. "Durable, AI-proof regulatory moat" is the category consensus adopted as personal conviction.
What you owe before an IC memo
- Net revenue retention, and share-of-clients penetration across the ~100 firms.
- The accuracy-over-volume curve: does cross-firm data measurably improve classification, or is each firm independent?
- A defensibility answer for the day Sage or A3 ships native Verifactu AI: what keeps a firm on Kabilio then?
the red team returns the thesis · { "verdict": null }
Yours to make, once you have the answers. The thesis is not dead. It is conditional on two facts you have not checked, and honest now about the fight with incumbents the consensus let you skip.
This example red-teams a bull case built on real El Foso evidence. It sharpens the thesis and returns it. It does not decide.