fleuret.ai

The core idea of replacing expensive manual pentests with an agentic system that delivers replayable PoCs and integrates with Jira is genuinely useful. Most founders pitch 'AI security' as vague scanners. This one at least targets the expensive consultancy gap with fixed pricing and EU data claims. Still, the execution risk is massive and the page never shows the actual AI doing novel exploitation chains.

The copy is drowning in startup clichés. 'Find, Prove, and Fix', 'PROOF NOT PROBABILITY', 'NOT A FIRM NOT A SCANNER A PLATFORM' – every line feels focus-grouped into meaninglessness. They repeat the same vulnerability list three times and lean on future-dated screenshots. The one honest section is the pricing comparison. The rest reads like someone discovered Notion templates and buzzword generators at the same time.

From the text it looks functional with some attempt at live workspace screenshots and a clean pricing table. However the repetitive scrolling lists of vulnerabilities and the awkward 'step 01 02 03' layout scream template. No real visual hierarchy, and the blog dates in 2026 already make the whole thing feel placeholder. It works as a brochure but nothing elevates it beyond average SaaS marketing site.

€4,000 with PoCs and zero false positives versus €15-30k consultancies is the only compelling claim on the page. The workspace-over-PDF angle and EU sovereignty story are real differentiators for compliance teams. Everything else collapses into generic 'agentic AI' handwaving. They never explain why their AI finds things scanners miss without hallucinating, which is the entire pitch.

The primary CTAs are 'Book a demo' and 'Get started' repeated everywhere with zero urgency or specificity. 'Free exposure check' exists but gets buried. The '0 findings, 0 invoice' guarantee is the strongest hook and they mostly hide it. Classic mistake of treating every visitor like they already understand pentesting budgets instead of forcing a clear next step.

It feels like a real product that might actually work, which is rare. The pricing honesty and workspace focus save it from total mediocrity. But the lazy copy, future-dated fake screenshots, and buzzword density make it look like every other AI security startup that raised too early and shipped too little. Competent enough to book a demo, not compelling enough to feel differentiated.