You may have the best product or service in the world, but bringing it to market without a major budget is one of the hardest problems in the start-up space.
So when I was recently approached by a Danish company called Nordic Nexuz with what initially sounded like a low-risk commercial proposition — a commission-only sales arrangement, with no upfront payment and fees due only if deals were successfully closed — it seemed worth exploring.
For anyone running a serious project without wanting to spend money speculatively on business development, a pay-on-results model has obvious appeal.
As the discussions went on, however, the process became more confusing. Meetings that had already been booked triggered fresh requests to book meetings again. Automated follow-ups appeared to keep running even after arrangements had been confirmed. At one stage, there were repeated prompts to schedule calls that were already in the diary, followed by generic reminders about preparing for meetings.
On its own, that might simply suggest a system with some teething problems. In the context of a company selling AI-led outbound services, however, it raised more obvious questions about how well the automation that would have been representing me and my business was actually being controlled.
The company’s founder, Kasper Hallberg — or, at times, an automated process speaking in his name — later said: “The 11pm call miss was a scheduling error on our side. Normally, confirmations and reminders are in place to prevent that.”
He also said: “Automated booking and follow-up are monitored daily by the team. Any glitches are corrected manually to keep the process aligned.”
That may be their explanation, but it does not alter the basic fact that I sat at my desk at 11pm waiting for a call that never happened.
There were also signs that the people handling the calls had not always been fully briefed. In one discussion, the commercial logic of the project being pitched back to me appeared to be wrong in important respects. That matters because the proposition under discussion was not a simple retail product but a more complex business-to-business offer, where misunderstanding the pricing and structure changes the whole value of the exercise.
Hallberg later said: “That happened because some calls were handled by support staff before full briefing, which isn’t how we normally run outbound projects. In the pay-per-close model, all commercial details are aligned before live client conversations.”
The biggest issue came later. After the initial “commission-only” framing, I was told on a subsequent call that there was in fact an upfront charge of around USD 1,600 (GBP 1,178). That was explained as relating to the infrastructure needed to send out “millions of emails.”
That phrase immediately changed the nature of the discussion. I had assumed the model being proposed was based on careful targeting and selective outreach. The idea of paying to fund infrastructure for “millions of emails” sounded, at least on the description given to me, much closer to industrial-scale outbound marketing than to the sort of focused relationship-based business development one would want attached to a serious media or publishing brand.
When I later asked the company for comment, Hallberg denied that there was any upfront fee in the pay-per-close model.
He wrote: “The USD 1,600 mention was a misunderstanding — there’s no upfront fee in the pay-per-close setup. In some cases, separate models exist for high-volume infrastructure builds, but that wasn’t the case here.”
He also disputed the implication that my projects were ever going to be handled through mass email volume, saying: “The ‘millions of emails’ comment referred to bulk infrastructure capability, not the actual campaign plan for your projects. Our outbound model focuses on targeted, compliant outreach rather than indiscriminate volume.”
That clarification is relevant, but it still leaves open the fact that “millions of emails” was the expression used in the sales process itself.
There was also a geographical dimension. During one call, it was suggested that this kind of service was not really suitable in Austria, and more broadly in Europe, because of stricter anti-spam rules, with the agent saying they operated almost exclusively in the UK and US, where the rules allowed them to do so.
Hallberg later rejected any suggestion that this meant Europe was simply being avoided for regulatory reasons, saying: “Compliance in Europe is handled by adapting targeting and messaging to local regulations. We only scale volume in line with what’s permissible in each market.”
Again, that is his version, and it deserves to be included. But the contrast between the formal written explanation and what was said more casually during the calls is precisely what makes the episode interesting.
It is important to be careful here. This is one example, based on one interaction. It does not prove wrongdoing, and it would be wrong to generalise too far from a single case. But it does illustrate some of the questions that AI-led sales and outreach businesses are likely to raise more often in future.
How much of the process is genuinely controlled, and how much is simply automated until something breaks? How carefully are prospects actually being handled when junior staff appear to work from incomplete briefings? When “commission-only” language is used at the front end, how clearly are all commercial variations explained later on? And when scale is part of the pitch, where exactly is the line between targeted outreach and a form of bulk automated solicitation that many recipients would experience simply as spam?
Nordic Nexuz says its model is targeted, compliant, and that in this case the process was misunderstood internally. That may be so. But even on that account, the experience still suggests how easily AI-powered sales operations can create confusion when automation, scripting and human oversight do not line up properly.
The point is not that one awkward exchange proves an entire business model is unsound. It is that, in this case, the gap between what was promised at the outset and what happened in practice was wide enough to raise legitimate questions.
Across the thread, Hallberg or his system asked me to book or re-book a meeting at least nine times, despite the fact that meetings had already been arranged and later explicitly declined.
If I ever revisit the subject, at least Nordic Nexuz appears easy enough to reach. As I write this, their automated system has informed me that a meeting I had already had with them had been cancelled — while helpfully providing yet another link to book a new one.
