Morgann Dawance

The through-line

Everything Morgann Dawance publishes circles one conviction: the tooling is extraordinary, and the tooling alone is worth nothing. "Le problème n'est pas l'IA ou l'automatisation. C'est un magnifique outil. Mais comme tout outil, il n'est efficace qu'à la hauteur de son artisan" [9]. The corollary runs through the rest of his thinking. Outbound technology "n'a jamais été aussi puissante" while commercial follow-up "n'a jamais été aussi éclaté" [1]; AI can reproduce a level of research at scale but cannot save a company that has no commercial traction to begin with [9]; and the industry's flirtation with fully automated prospecting was always a fantasy. When Artisan, the AI BDR company behind the "Stop hiring humans" slogan and a $2M marketing budget to broadcast it, publicly reversed course and started recruiting a human BDR, his reaction was not schadenfreude but relief: "ce qui a échoué chez Artisan, ce n'est pas l'IA" [5], and the market is finally abandoning "les formules magiques à deux balles" [5].

The position has not so much shifted as deepened. The 2023 framing was mechanical and durable: three pillars, all load-bearing [14][15]. The later work adds a layer on top, where AI and agentic workflows change how the work gets done without changing what makes it work, and where the missing piece is no longer capability but usability for the people who actually sell.

On the three pillars of prospecting

His base model of a prospecting campaign has three legs, and he insists on the stool metaphor deliberately: remove one and the whole thing falls over [14][15]. The first is technical setup and data quality, meaning emails properly configured, DNS settings correct, so that messages land in the main inbox rather than in spam [14][15]. The second is targeting: identifying the right contacts, the ones whose needs genuinely align with the value proposition [14][15]. The third is messaging, copy that resonates with the audience it is aimed at [14][15]. The logic linking them is what matters. Careful technical work exists so that carefully crafted messages actually arrive [14], and when all three are aligned, prospecting delivers exceptional results [15].

On AI as a craft rather than a miracle

Asked during a training session for the hub.brussels teams whether AI tools can generate genuinely qualitative leads, he refused the binary answer: "ça dépend. Oui, pour la masse. Absolument pas, lorsqu'on sait où et pourquoi l'utiliser" [9]. Behind that question he hears a deeper fear, that AI will dehumanise exchanges and make prospecting ineffective [9]. His answer is that the real error is expecting the plug-in to be miraculous, expecting an automation or an AI to rescue a business that has not found commercial traction [9]. What AI does deliver is the ability to reproduce a level of research at a scale that was previously impossible [9]. That is also the model Leadix has been running for four years and the one Artisan arrived at late, framed by its CEO Jaspar Carmichael-Jack as helping humans become more effective, a "Super BDR" [5].

He is equally impatient with the discourse on the other side. Feeds full of "l'IA va remplacer des équipes entières" leave people nodding along while sensing they are missing something, and nobody shows them concretely how to do it [11]. His response was to spend "des milliers d'heures à poncer l'IA" across Claude, ChatGPT, Cowork and Claude Code, testing, breaking and iterating, then compile the result into a guide aimed explicitly at people who have not had that luxury of research time [11]. The stated goal is to turn Claude into "ton premier employé virtuel", with no prior knowledge required [11].

On the four traps that ruin your prompts

He attributes most people's frustration with LLMs to four mechanics nobody explains [8][11]. Hallucination comes first: the model invents with total confidence, because it is a predictive model of what is most "logique", which means output can look perfectly coherent and be entirely false. Figures, names, dates and quotations should be verified always, every time [8]. Second is the knowledge cut-off, the training end date, which he compares to someone disconnected from all information for a long stretch; for recent information the answer is web search or connectors, not the model itself [8]. Third is the context window, since the model does not remember everything and per-conversation memory is short [8]. Alongside the traps he teaches three distinct ways of using Claude, noting that most people know only one, and which model to use when [11], plus a method for prompting to structure a task and get the expected output [10].

On agentic GTM and the terminal

His read on where go-to-market work is moving is blunt: "Une grosse partie des workflows GTM ne se construisent plus dans une interface. Ils se construisent dans Claude Code, de manière agentique" [7]. Sourcing, enrichment, scoring and CRM routing now run through a terminal, and the one conspicuous absentee from that stack was the central integrator, Clay, which made its arrival in the terminal heavily anticipated [7]. After several hours digging into the integration his verdict on the positive side is that CLAY CLI builds simple workflows faster without touching the app: sourcing 200 fintech CFOs in BENELUX, waterfall enrichment of email and phone from a CSV lead list, quick lead scoring, all built autonomously thanks to direct access to more than 200 providers [7]. This is the daily working reality he describes elsewhere, with teams operating inside Lemlist, Clay, N8N and Claude Code [1].

On the dashboard that was missing

The same tools he praises carry a design flaw he keeps returning to: "ils sont pensés pour ceux qui construisent, ceux qui pensent en DNS, webhook, API" [1]. That is not who uses the output. "Un Head of Sales, un lundi à 8h, a une seule question : 'où en est mon pipeline ?'" and a salesperson wants to focus on prospects, not on the technology behind them [1]. So Leadix built the bridge between powerful outbound tech and fragmented commercial follow-up: a genuine steering dashboard where sales teams track and convert their best leads, tag colleagues, and leave themselves notes and reminders so no opportunity is missed [1]. It is now deployed across the entire client base [1].

On pricing, packaging and value perception

Doubling revenue without doubling lead volume is possible, and he says so without hedging, though the method is not trivial [4]. The centrepiece is the pricing roadmap, illustrated by Apple: the first iPhone launched in 2007 at $499, a price considered abnormal at the time, and experts were unanimous that prices would fall. Nineteen years later paying $1500+ for an iPhone is normal [3]. That evolution was not accidental. It was a pricing roadmap put in place from the first launch, where every keynote, every announcement and every feature served a precise objective of value perception, and the adapted price followed [3]. The rest of the territory he maps with Laurent-David Hostyn of Givemefive covers the anatomy of good packaging, how to build an offer that captures maximum value, the signals that betray broken packaging, the ideal number of packs for a company, when displaying prices is far more destructive than helpful, the pricing error too many companies make, and versioning as a way to raise prices without ever announcing an increase [4].

On copy and the human link

However much AI takes over the process, he holds the line on writing: "L'IA a beau prendre de plus en plus de place dans nos process, nous restons convaincus qu'un bon copy fait la différence pour créer du lien" [13]. Leadix he defines in two words as a B2B go-to-market agency that introduces clients to their future clients, "un bon message à la fois" [13]. The profile he hires against is specific: confirmed B2B copywriting experience, ideally in outbound and cold email; someone who grasps commercial stakes quickly and masters sales psychology, with a growth or GTM background a strong plus; who writes straight to the point without jargon or unnecessary words; who is curious by nature and constantly digs for new angles; bilingual FR/EN with Dutch as the jackpot; and autonomous enough to learn fast and self-edit without anyone reworking behind them [13]. Applications do not require a cover letter [13].

On client relationships and under-promising

Being paid for a service is the normal contract, and he says so plainly: accompanying clients in their customer acquisition is the job, and being remunerated for it is the counterpart [6]. What moves him is what falls outside that contract. Gifts from partners are one thing, but when clients go the extra mile with invitations to events, thank-you gifts or a handwritten letter, "c'est la plus belle des sensations au monde" [6]. He frames it as reciprocity of a principle he has always applied: "J'ai toujours eu la volonté d'under-promise et over-deliver pour l'ensemble de nos clients", and receiving the same treatment back was a genuine surprise [6].

On celebrating the milestone instead of ticking the box

He names his own failing without softening it: "Je ne fête pas assez. C'est un de mes défauts" [2]. The pattern is to hit a target, glance at the KPI, think "Top", and be onto the next thing thirty seconds later, which over time cuts you off from the most interesting part, enjoying the journey [2]. With Leadix having gained "5 ans de maturité en 5 mois", he almost repeated the reflex and instead stopped to mark the moment with Diego [2]. The list he stopped for: developing and onboarding clients onto the new Leadix app, a complete modernisation of internal processes, recruiting a new profile, beating the monthly revenue record, rethinking their AI-enabled services, and above all becoming associates [2]. They deliberately did not want a simple team-building exercise [2].

Takeaways

• Treat prospecting as a three-legged stool: technical setup and deliverability, targeting whose needs match your value proposition, and messaging that resonates. All three must hold or none does [14][15].
• Never expect automation or an AI layer to rescue a business without commercial traction. AI reproduces research at scale, and it is only as effective as the person wielding it [9].
• Verify every figure, name, date and quotation an LLM produces, use web search or connectors for anything recent rather than the model itself, and account for a short per-conversation context window [8].
• GTM workflows are moving out of interfaces and into the terminal; CLAY CLI now handles sourcing, waterfall enrichment and lead scoring autonomously across 200+ providers, without opening the app [7].
• Build for the Monday-8am question, "où en est mon pipeline ?", not for the person who thinks in DNS, webhooks and APIs. Sales teams need tracking, tagging, notes and reminders, not plumbing [1].
• Give your company a pricing roadmap: Apple moved the iPhone from $499 to $1500+ by engineering value perception through every keynote and feature, not by accident [3].
• Hire copywriters who write straight to the point without jargon, grasp sales psychology fast, and self-edit; good copy remains the differentiator for creating connection even as AI expands in the process [13].
• Under-promise and over-deliver as the operating default, and stop long enough to actually mark the wins instead of ticking the KPI and moving on [6][2].