published by policy, not bravado
Reasons not to hire us.
Before launch, we ran adversarial reviews against our own business: a skeptical buyer, a competitor analyst, and a solo-operator realist scored the thesis and tried to kill it. The strongest objections are below, unedited in substance, with our answer where we have one and a concession where we do not. If a vendor has never done this exercise, ask them why.
The objections
"You are one person. What happens when you are sick, busy, or gone?"
Concession first: this is true and it caps us at roughly three concurrent engagements. Our answer: every engagement is designed to survive us — fixed scope, named written artifacts, your infrastructure, your keys, and a handover test where your engineer redeploys without us in the room. The autopsy format exists precisely because it is small enough for one senior person to deliver excellently. If you need a bench of twenty consultants, we are the wrong vendor and will say so on the call.
"The buyer who approved the dead pilot has to admit it died to hire you. That's politically radioactive."
Concession: this is our hardest sales problem, identified by our own buyer-lens reviewer. Our answer: the autopsy is framed as an asset recovery, not a blame exercise — the report separates technical causes from organizational ones, and the readout gives the sponsor a costed path forward they get credit for. But yes: if your politics cannot survive the word "autopsy", the engagement will stall. We would rather lose the deal than rename the product.
"Any competent freelancer can copy 'senior engineer fixes your AI' tomorrow. Where's the moat?"
Concession: the positioning is copyable; our competitor-lens reviewer scored our moat 3.7/10 and the commoditization clock at 12–24 months. Our answer: the durable layer is not the plumbing — it is the judgment (risk classification rationale, oversight design, eval design) plus the compounding template library, plus an EU legal entity with a decade of history that a DPO can approve. Also: most freelancers will not publish their red team. You are reading ours.
"Zapier MCP has 9,000 pre-authenticated integrations. Composio raised $29M for the same. Why pay you for integration?"
Agreed — and we would use those exact products in your build where they fit. Our answer: nobody's stalled pilot died because a connector was missing. They die at accuracy, adoption, ownership, and audit-readiness — the parts after the plumbing. If your problem genuinely is a missing connector, buy the product, not us. The self-diagnostic will tell you which case you are.
"The AI Act deadline slipped to December 2027. Your compliance urgency is manufactured."
Correct, and we publish the slippage ourselves. Our answer: we do not sell deadline panic. The compliance artifacts we ship — inventory, DPIA-lite, oversight design, logging — are the same documents that make a system operable and auditable regardless of which year enforcement lands. If a vendor is selling you August-2026 fear, show them the Travers Smith briefing.
"Clients believe AI lets them do this in-house now. 33% say exactly that."
Some can, and the honest version of our pitch says so. Our answer: the same survey wave shows the failure inventory growing — 42% of companies abandoned most AI initiatives. In-housing works when you have a senior engineer with production LLM experience and spare capacity. If you have that person, keep them and skip us. Most 25–250-person firms do not have that person; renting them for eight days costs €2,500.
"Your prices are suspiciously low for 'senior'. €2,500 for eight days is ~€64/hour."
Sharp catch — our own financial model says the same. Our answer: the autopsy is deliberately priced as customer acquisition, not margin. It is a paid, useful sales call. The margin lives in the Build (~€165/hr effective) and the retainer. We would rather tell you that than pretend the entry price is charity.
"No case studies. Zero references. Why would I be first?"
Concession: at launch this is simply true. Our answer: the first two or three Builds are priced as discounted reference engagements, and the specimen autopsy report is public so you can judge the actual work product before paying. The company itself is not new — Coherent Perspective has operated since 2016 — but this productized offer is, and pretending otherwise would violate our own first principle.
the full record
The complete adversarial review — four business theses scored by three independent reviewers, kill votes included, plus the objections that killed two of the theses — is preserved in the company build log. Ask and we will send it. This page carries the objections that survived into the business we actually launched.