Partners have the customers and the catalogue. What they do not have is presales.

Cyberangels was founded in Milan in 2021 and spent its first years inside the insurance industry. We worked in open innovation with Vittoria Hub, the insurtech incubator of Vittoria Assicurazioni, one of our investors, and the product was a self-assessment that told a small company which risks it actually carried and what cover those risks needed. That October we won a call for solutions that Poste Italiane had opened to insurtech startups, and the prize was a three-month trial with Poste Italiane and Poste Assicura. We entered the insurtech arena and came first in it.

Recognitions: winner, SME Insurtech – Call for Solutions, Poste Italiane with LVenture Group, 2021; best insurtech product, 2023; finalist, TIM startup contest, 2024.

None of it fixed the buyer. Small companies at the far end of the chain were not purchasing security on their own: they agreed they needed it, asked sensible questions, then did not sign. We had won the arena and the signature still did not come. Our summary of those years is blunt. Cybersecurity for small and medium enterprises is not something everyone wants to buy.

The first hint of an answer came from a bundle that had gone cold. TIM had a security offer in the market that was barely moving and asked us to relaunch it. Technically we changed almost nothing. What changed was who the product was for: not the small business at the end of the line, but the specialist selling to it, who now had a reason to call and a service to attach. The bundle started selling again, and we did not yet understand that this was the whole company.

Two years with little to show in public. We built, and we did research with Italian universities and hospitals. Revenue grew each year and the loss narrowed to almost nothing, most of what remained being the amortisation of software we had written ourselves. A small company, moving the right way.

Two weeks of demos at MWC in Barcelona. The question that kept coming back was not whether the AI could find the risk; everyone assumed it could. It was whether what it suggested would mean anything to that operator's own customers. Relevance, not detection.

That spring we read through about forty vendors in AI presales, close enough to our own work to be uncomfortable, and decided not to pivot. The scan said the direction was right. It also said we were describing it wrongly.

So we changed one word. Not a cybersecurity platform: a platform for the people selling cybersecurity.

Cybersecurity became the object of the sentence instead of the subject: what the partner sells, not what we are. We adopted the claim knowing the product was about seventy per cent of the way to keeping it, and wrote that down at the time. By 20 July the language had settled into the line we still use: “Automate cybersecurity presales. Copilot the sale.” In August we cleared out what no longer fitted, old product names and old price tiers included, and kept one brand and one unit of price.

The product is two co-workers, one that prepares the call and one that stays with the rep to the close. Over 50 partners and SMEs use it, and on screen it carries the partner's brand rather than ours.

The scanning, the analysis and the pitch are ours. Everything the customer sees belongs to the partner.