AI in Sales: Magnet World Quadruples Demos

Case Study · 13. Juli 2026 · Leon J. Hermann

AI in sales for SMEs: Read how Magnet World quadrupled demos, sharpened its pipeline, and gained sales capacity without new SDRs. Check it out now.

You constantly hear on LinkedIn that AI in sales is primarily about volume: more contacts, more emails. That's only half true. For Magnet World, a specialist in industrial magnets, the opposite was the lever. Less wastage. More precise mathematics. And a sales team that didn't need another list of 8,000 semi-suitable companies, but finally knew which 400 accounts really mattered.

I'm writing this case study based on my work at Amplifa. Not as an outside observer. Magnet World is exactly the type of SME where classic sales tech promises often sound like they're from another industry: too loud, too broad, too far removed from the reality of workpieces, load profiles, safety factors, and purchasing rounds where a technical manager only listens once someone understands the application.

The short version: After implementing Amplifa, Magnet World quadrupled its qualified outbound demos. Not because AI suddenly sold magically. That's not entirely true. It's because the company cleanly coded its Ideal Customer Profile, systematically expanded its account coverage, and tied its initial outreach so closely to real use cases that cold outreach once again became a sensible business tool.

AI in Sales at Magnet World: Why now?

Magnet World doesn't sell software you test with a credit card. Magnet World sells industrial magnets and magnetic systems for manufacturing, handling, automation, and special machine construction. Lifting magnets, clamping systems, magnetic assemblies, customized solutions. Products where a wrong fit doesn't just lead to cancellation, but in the worst case, to downtime, scrap, or a safety risk.

That makes sales complicated. A buyer at a steel service center in Duisburg reacts differently than a development manager at an automation OEM in East Westphalia. A production manager in a recycling plant thinks in terms of scrap throughput, crane cycles, and maintenance. A design engineer at a machine builder thinks in terms of installation space, temperature, cycle time, and interfaces. All can be customers. But not with the same story.

In March 2025, we analyzed the initial situation with Tobias, Sales Manager at Magnet World in Bielefeld. The smell of cutting fluid still clung to his jacket when he came from production to the workshop (you remember such details when you spend all day looking at CRM fields). His statement was dry: "We have enough market. We just don't have a clean machine to work it."

That's the point. Many managing directors in the manufacturing SME sector believe their pipeline problem is demand. In my experience, it's more often precision. Companies often know very well whom they can help. But this knowledge is stored in people's heads, old quote folders, Excel lists, trade fair notes, and CRM free text fields. No system can reliably prioritize from this. No SDR can build a scalable outbound machine from this in 90 days.

At the same time, pressure is increasing. Magnet-Schultz Memmingen reported €364 million in sales and around 11 percent growth for its 114th fiscal year, driven by automotive, hydraulics, and commercial vehicle projects, among others. This is not a direct benchmark for Magnet World, of course. Magnet-Schultz is much larger. But it shows where the market is heading: magnetics are found in mobility, hydraulics, energy, automation, and technically demanding niches.

And while demand is rising in individual segments, sales capacities remain scarce. An experienced technical salesperson for industrial magnets cannot be hired in six weeks. Honestly? In some regions, you can't find one at all. If existing key accounts also need to be managed, quotes calculated, inquiries from design answered, and trade fairs prepared, what's left for systematic new business is what still has energy after 5 PM. So, not much.

Context: Industrial Magnets are a Niche, but not a Small Niche

From the outside, the market seems inconspicuous. A magnet is a magnet, one quickly thinks. Not true. Worlds lie between a permanent magnet for a simple component, an electromagnetic lifting magnet for steel plates, and a magnetic assembly in an actuator. Different materials. Different safety requirements. Different purchasing logic.

In discussions with manufacturing companies, I repeatedly see the same mistake: sales teams segment by industry, but not by application. "Mechanical engineering" is not an ICP. That's a continent. For Magnet World, the better signals were much more specific: steel plate handling, CNC clamping technology, crane systems, recycling lines, hydraulic assemblies, OEMs with magnetic actuators, automation integrators with gripping technology projects.

An example. A medium-sized machine builder from Baden-Württemberg can be completely irrelevant if they only assemble plastic housings. A small company with 80 employees near Hagen, on the other hand, can be highly interesting if heavy sheets are moved there every day, crane hooks clatter, and one welding table after another is occupied. Company size alone is of little help here. Webasto, Brose, Schaeffler, or Festo immediately appear in databases. The real opportunities in SMEs are often hidden in less polished company profiles.

Added to this is raw material pressure. The German Raw Materials Agency has been observing critical raw materials for future technologies for years, and rare earths for high-performance magnets are part of this debate. Neodymium, dysprosium, terbium – these are not terms you should explain in a generic cold email. But they influence prices, availability, and technical advice. Anyone who only writes "We supply high-quality magnets" in sales has already lost.

Why Classic Lead Generation Wasn't Enough for Magnet World

Before Amplifa, Magnet World didn't have an empty CRM. That would have been easy. There were existing customers, old trade fair contacts, quote histories, contacts from associations, commercial register lists, manually researched companies, and a few purchased datasets that looked decent at first glance and smelled like work at second glance. The problem wasn't a lack of data. It was data mixing.

When a sales manager sees 2,300 accounts in the CRM, it feels like market coverage. Until you ask: Which of them really have lifting processes with ferromagnetic loads? Which are in a current investment phase? Which are OEMs and which are dealers? Which locations have production and which only administration? Which contacts decide on safety, cycle time, or procurement? Then the perceived pipeline shrinks pretty quickly.

Andrea, Head of Sales at a hidden champion in Bielefeld, recently told me in another project: "Our list was big enough to reassure us, but too poor to plan sales." That pretty much describes Magnet World. A large list in B2B sales is often a sedative. Not a control instrument.

The second weakness was in the approach. A sales employee could write good messages if they had time. But they rarely did. So sequences were created that were technically correct but interchangeable. "We are specialists in industrial magnets." "We support companies with safe lifting processes." "We would be happy to present our solutions to you." None of this is wrong. It just doesn't create movement.

At Magnet World, the monthly number of qualified demos from active outbound was in the low single digits before the project. Internally, it fluctuated depending on trade fairs, holidays, quote load, and market mood. After six months with Amplifa, it was consistently four times higher. Important: We're talking about qualified demos with a technical fit, not calendar entries with students, consultants, or people looking for a magnet for their refrigerator.

Detailed Analysis Part 1: The ICP Was the Product

The most important part of the project wasn't the first email. It was the ICP. Sounds boring. But it's the part where SMEs either make money or burn out their sales team.

At Magnet World, we didn't work with "target group mechanical engineering DACH." Too broad. Instead, we defined four core segments, each with its own data criteria, triggers, and exclusion logic. Steel service and metal trading. Recycling and scrap handling. Mechanical and plant engineering with clamping or gripping processes. OEMs in hydraulics, actuators, and automation, where magnetic components can be part of a larger system.

Then came the uncomfortable part: throwing out accounts. Many teams love to inflate the TAM. Total Addressable Market sounds better when it's big. But for Magnet World, clarity was more valuable than size. An account without ferromagnetic loads, without relevant manufacturing, or without a technical decision structure was not prioritized. Done.

We combined public company data, industry classifications, website signals, job advertisements, product pages, location information, and CRM history. If a company mentions flame cutting machines, steel plate cutting, crane runways, or heavy CNC machining on its website, that's a different signal than "metal processing" in a database. If job ads are looking for maintenance technicians for crane systems, that's no coincidence. That's sales information.

What we specifically see at Amplifa: For technical niche providers in DACH SMEs, after the first ICP cleanup, usually 35 to 55 percent of seemingly suitable accounts are removed from prioritization. At Magnet World, the value in the first cleanup was 47 percent. That sounds harsh. But it was the reason why the remaining sequences performed. Bad accounts don't just cause bad numbers, they wear out good salespeople.

After that, we modeled the buying committees. Not just purchasing. For industrial magnets, four roles are often relevant: production or plant management for throughput and ergonomics, design or engineering for integration, occupational safety for risk and standards, purchasing for supplier approval and price. For larger companies like Schaeffler or Brose, location logic is added. Central purchasing is not necessarily where the lifting problem arises.

This changed the messages. A production manager didn't get text about material classes. They got an introduction about load changes, crane time, and manual slings. A design engineer didn't get a product catalog. They got a hypothesis about installation space, gripping situation, or integration into a line. A buyer only later received information about delivery capability, documentation, and variant logic. Order matters.

"The surprising thing wasn't that AI could write texts. The surprising thing was how quickly we saw which accounts we had overestimated for years."

— Tobias, Sales Manager at Magnet World, Bielefeld

Tobias hit a sore spot with that. Many sales teams measure activity. Few measure the quality of account selection rigorously enough. If an employee contacts 40 companies and gets no response, the subject line is often tweaked. Sometimes, however, one should ask whether 27 of those 40 companies even have a problem that the provider can solve.

At Magnet World, we therefore didn't just look at open or reply rates. These can be misleading. We measured per segment: proportion of positive responses, proportion of technical inquiries, proportion of qualified demos, proportion of demos with an identified project or reliable need, average time to the first meaningful reaction, and proportion of accounts that went into a nurture workflow. This is less pretty than a single metric. But it provides better control.

Counter-argument: Why not just hire more SDRs?

This question almost always comes up. And it's legitimate. If outbound generates too few demos, why not hire two junior SDRs, give them LinkedIn Sales Navigator, and off they go? Because it rarely works in tight technical markets. Well, almost. It works in the short term as an activity booster. But not as a system.

A junior SDR can learn to send a sequence. They can also learn to distinguish a steel service from a toolmaker. But they cannot reliably recognize whether a company needs lifting magnets, magnetic clamping technology, or customized magnetic assemblies. They don't know whether an objection from purchasing is commercial or technical. And if they have to have every second lead checked internally by the Senior Sales Engineer, you haven't created capacity. You've relabeled bottleneck work.

This was particularly evident at Magnet World. The sales team consisted of experienced people who understood customer applications. These were precisely the people you didn't want to block with initial lists, address verification, and follow-up reminders. Their time belongs in technical discussions, quote clarification, and co-development. Not in asking whether Mr. Müller is still with the company or whether a GmbH in Saxony actually produces.

I'm pretty blunt here: Anyone who wants to scale in technical B2B sales in 2026 solely through more headcount is building an expensive coordination problem. Especially in SMEs. The better question is not "human or AI?". The better question is: Which work deserves human experience, and which work must a system reliably perform every day?

The answer at Magnet World was clear. People for technical qualification, trust-building, and quote strategy. Amplifa for account research, ICP scoring, buying committee mapping, personalized sequences, follow-ups, and learning loops. That sounds less romantic than "AI sells autonomously." But it's closer to revenue.

What Amplifa Specifically Did: Modules, Workflows, Learning Loops

I don't like case studies that stop at "we used AI." That doesn't help anyone. So, specifically: At Magnet World, the setup consisted of five operational building blocks. ICP playbooks. Account Universe Build. Contact and Committee Mapping. Sequencing with technical personalization. Performance review with segment decisions.

In the first step, we analyzed the best existing customers and lost deals. Not just by revenue. By application. A lost deal can be more valuable than a won one if it shows why a segment doesn't fit. At Magnet World, for example, there were accounts that moved large quantities of metal, but due to existing crane and safety logic, were unlikely to switch in the short term. Others seemed smaller but had an acute bottleneck in manual handling. These differences belong in the ICP, not in gut feeling.

Then we built an account landscape for DACH and selected EU markets. The focus was initially on Germany, Austria, and Switzerland, because sales, service, and technical consulting could react fastest there. After that came neighboring markets with similar industrial and linguistic logic. No world map. No "Let's try the USA." SMEs often get bogged down with internationalization before the domestic market is cleanly covered.

The data basis was enriched per account: industry, location type, production notes, machine or process signals, potential application, contact roles, existing CRM history, triggers from website or job ads, exclusion reasons. An account without a clear reason was not given high priority. This was unusual at first. Sales loves opportunities. Operations loves criteria. In this project, operations had to win.

For the outreach, we didn't build a one-size-fits-all sequence. Steel service centers received a different narrative than hydraulic OEMs. For recycling companies, the operational bottleneck was in the foreground. For automation integrators, it was more about project capability and integration. For OEMs, it was about specification, material questions, and series logic. The tone was technical, concise, and without marketing gloss. A production manager notices after two lines whether you know their world.

A detail that is often underestimated: follow-up. Many sales teams send a good first message and then break off. Not out of laziness. Because everyday life gets in the way. Quotes, escalations, internal meetings, trade fair preparation, a customer calls about a drawing. At Magnet World, Amplifa took over the consistent, segment-logical follow-up. Not annoying. But present.

The learning loops ran weekly. Which segments respond? Which roles react? Where is a "no need" honest, where is it just timing? Which message generates technical inquiries instead of polite responses? Which accounts belong in nurture because a project will only be relevant in the next budget window? Sales here is like manufacturing: If you don't measure, you optimize by noise.

Amplifa ICP Playbook The playbook shows how technical B2B providers segment their target customers not by gut feeling, but by purchase-relevant signals.

Industry Comparison: What Other SMEs Do Differently

I see three patterns in the manufacturing SME sector. First: large brands with strong inbound structures. Trumpf, Festo, Phoenix Contact – these companies have visibility, content, partner networks, trade fairs, international sales organizations. Even there, outbound is becoming more important, but they are not starting from scratch. A niche provider like Magnet World does not have this air cover. If no one googles your specific magnet module, SEO alone won't help you.

Second: Hidden champions with a good reputation in a narrow customer base. This was closer to Magnet World. You are known by those who know you. The sentence sounds trivial, but it is dangerous. Because outside the established customer circles, the company often hardly exists. At the trade fair, someone remembers the name, but in the purchasing system of the next plant, a competitor appears because they asked earlier.

Third: Providers with commodity pressure. There, sales are quickly driven by price. While industrial magnets are partly comparable, the real margins lie in application, safety, integration, and consulting. Those who enter the buying process too late end up in the price column. Those who come early enough with a plausible technical hypothesis can set the framework. That's not a nice saying. That decides the contribution margin.

A CSO from a machine builder in Nuremberg told me three weeks ago: "That doesn't work for us, our customers only buy through existing contacts." I never immediately believe that sentence. Usually, it means: We haven't yet found a way to prove relevance to new contacts quickly enough. Existing contacts are valuable. But they are not a market. They are a segment.

Magnet World had to enlarge exactly this segment. Not through noise. Through relevance. The company had technical substance. Amplifa had to translate this substance into repeatable market cultivation.

Practical Example: From 3 Demos to 12 Qualified Demos Per Month

The number "fourfold" sounds round. But there was no linear progression behind it. In the first month, much was diagnostic. Old data, unclear roles, duplicates, companies without production sites, contacts who had long since changed jobs. A CRM doesn't smell of oil or metal dust, but sometimes it feels exactly the same: You notice where work has been done and where something has just been filed away.

Before the project, Magnet World generated an average of about three qualified outbound demos per month. Depending on the month, it was two, sometimes four. After six months, the stable average was twelve qualified demos per month. Not every appointment became an opportunity. Of course not. But the pipeline became more predictable, and that was almost more important to Tobias than the peak.

The appointment rate increased primarily in two segments: steel service and automation integrators. In steel service, messages addressing specific lifting and safety processes worked. For integrators, the question of recurring gripping or clamping problems in customer projects worked. That sounds small. It's not. A good outbound message doesn't sell the product. It sells the assumption that a conversation saves time.

A concrete workflow looked like this: Amplifa identified a metal processor with indications of sheet metal cutting, crane runways, and heavy profile steel. The account received a high fit score. Production management, maintenance, and purchasing were marked as relevant roles. The first message to production management referred to the reduction of manual slinging operations and the question of whether certain load profiles are moved regularly. The follow-up to maintenance addressed maintenance and risk of failure. Only later did a technical appointment come into play.

The result was not Hollywood. A response after nine days. Then a brief exchange. Then an appointment with production and purchasing. Then technical review. That's exactly what B2B looks like. Not glamorous, but measurable.

The most important side effect: Sales talked differently internally about market opportunities. Previously, it was often said: "We should do more in recycling." Afterwards, it was said: "Recycling companies with scrap cranes and their own maintenance respond to throughput and wear arguments, but smaller dealers without transshipment technology are out." That's a different sentence. A better sentence.

AI in Sales Does Not Mean Autopilot

I'm not a fan of autopilot rhetoric. It sells well. It disappoints better. At Magnet World, Amplifa was not a substitute for sales, but an operating system for market cultivation. The salespeople remained crucial. They just had to spend less time on poor preparatory work.

This is particularly evident with technical objections. If a prospect asks whether a magnetic solution works at certain temperatures, load cycles, or surfaces, no AI should improvise. Then a specialist must step in. Period. But AI can ensure that this specialist speaks to the right people more often.

This is precisely where the ROI lies. Not in saved people. In better utilized experience. A Senior Sales Engineer who spends five hours less per week on unsuitable leads and instead conducts two additional technical discussions changes the revenue leverage more than any subject line optimization.

In April 2025, we discussed the first segment figures with Magnet World. One segment looked strong on paper but responded weakly. Previously, one might have continued because the industry was "strategic." We deprioritized it. Another segment was smaller but reacted with concrete project questions. That got more volume. So simple. So rare.

Why Timing is Underestimated in Technical Sales

Many decision-makers don't buy industrial magnets because an email is nicely worded. They buy when a project, a safety problem, a capacity bottleneck, or a replacement investment is imminent. Outbound must therefore not only convince but also make timing visible.

At Magnet World, we built nurture logics for this. Not every account with a fit is immediately pressured. Some belong in a later cycle, for example, when a new line is announced, when job advertisements indicate an expansion of maintenance, or when a trade fair like EMO, automatica, or LogiMAT is approaching. The difference between "no interest" and "not now" is real money in SMEs.

An example from another Amplifa project with a supplier near Stuttgart: An account responded in the first contact only with "contact us in autumn." Previously, that would have been lost in the CRM. With proper nurture, an appointment came about six months later because a new production line was being planned in the meantime. Was that AI magic? No. It was discipline that a system maintains better than people in day-to-day business.

Amplifa Product Amplifa combines ICP scoring, account research, AI SDR workflows, and performance learning for B2B sales teams in SMEs.

FAQ: When is AI in sales worthwhile for niche providers?

AI in sales is worthwhile for niche providers when three conditions are met: the target market is large enough to be systematically worked; the ICP can be identified by real signals; and the sales team has expertise that can be translated into good conversation management. If these three things are missing, even the best platform will only generate activity.

FAQ: What data does an SME need before starting?

Not perfect data. But usable truth. At Magnet World, existing customer lists, lost quotes, CRM history, sales segment knowledge, product applications, and clear exclusion criteria were sufficient to start. Perfection is often an excuse. Those who wait until the CRM is clean sometimes wait until the next ERP migration.

FAQ: Does an AI SDR replace technical sales?

No. And anyone who promises that has probably never sold a technical purchasing process. An AI SDR can scale research, prioritization, initial contact, and follow-up. Technical evaluation, trust, specification, risk assessment, and quoting remain human work. At Magnet World, precisely this separation was the reason why the setup worked.

The 7 Steps Other SMEs Can Adopt

If I transfer the Magnet World implementation to other manufacturing SMEs, seven steps remain. Not as theory. As a work plan.

1. Analyze won and lost deals by application, not just by industry or revenue. The question is: What specific problem was solved or not solved?
2. Build ICP segments with exclusion criteria. A good ICP not only says who you address, but also who you deliberately ignore.
3. Combine account signals from multiple sources: website, job ads, locations, product pages, CRM history, industry classifications, and public indications of machines or processes.
4. Define buying committees per segment. Production, engineering, purchasing, quality, and safety do not respond to the same message.
5. Write sequences by use case, not by product category. A lifting magnet is a throughput issue for a production manager, a risk issue for occupational safety, and a supplier issue for purchasing.
6. Measure weekly by segment. Not just response rate, but demo quality, technical inquiries, opportunity share, and reasons for rejections.
7. Protect human expertise. Senior Sales and Engineering belong in qualified conversations, not in list hygiene and initial contact routines.

The most uncomfortable step is number two. Exclusion criteria feel like throwing away revenue. But you're not. You're throwing away distraction. That's sometimes the scarcest resource in SMEs.

ROI Perspective: What Four Times More Demos Really Mean

Four times more demos sound good. But managing directors shouldn't stop at demos. A demo is not revenue. It is a qualified entry into a buying process. The ROI only arises when it leads to more valid opportunities, better forecasts, and more won projects.

At Magnet World, the first measurable effect was the relief of the sales team. Less time spent on unsuitable companies. More conversations with technical substance. The second effect was pipeline transparency. Suddenly, Tobias could not only say that outbound was "running," but also which segments were contributing what. The third effect was strategic: The company saw white space that it had only suspected before.

I prefer to avoid fantasy ROI with false precision. No one should claim that a single workflow guarantees x million euros in sales. But one can certainly calculate what a doubling or quadrupling of qualified initial conversations means with a stable opportunity rate. If a technical provider with project business generates only two real opportunities from twelve qualified demos per month and wins one of them over time, that changes annual planning.

Three Learnings from Magnet World

Learning 1: Narrow Niches Need More Math, Not More Volume

Magnet World didn't win because the outreach got louder. It got more precise. That's a difference many teams underestimate. The narrower the niche, the less volume without fit brings. In broad markets, you can sometimes buy your way out of wastage. In technical SMEs, it becomes expensive because every bad inquiry eats up internal expert time.

The math started with account fit. Which signals correlate with need? Which industry designations are misleading? Which locations are relevant? Which roles respond? Which triggers indicate timing? That sounds like analytical work. It is. But precisely this work determines whether AI in sales prepares revenue or merely generates activity.

Learning 2: Technical Personalization is Not a Nice Addition

Many personalizations in outbound are embarrassing. "I saw you've been with the company for 17 years." Nice. So what? At Magnet World, personalization had to be process-related. A reference to steel plate handling, crane cycles, magnetic clamping technology, or integration issues is relevant. A LinkedIn compliment is not.

The best responses didn't come from the nicest texts, but from the most plausible hypotheses. "We often see bottlenecks in manual slinging operations in companies with similar cutting processes" is stronger than "We are a leading provider." No one cares about being leading as long as their own problem remains unsolved.

Learning 3: Pipeline Management Starts Before the First Meeting

Many sales teams only start pipeline management when an opportunity is in the CRM. Too late. At Magnet World, pipeline management began with the question of which accounts even get a systematic place in the market approach. If this layer is unclean, the forecast will later only be a nice report about old mistakes.

After implementation, Magnet World could see per segment where outbound was working, where nurture made sense, and where it was better to stop. Stopping is an underestimated sales skill. If you follow up on everything, you end up not following up on anything properly.

What This Case Study Does Not Prove

It doesn't prove that every company quadruples its demos with Amplifa. That would be unserious. Magnet World had several prerequisites: a clear technical offering, genuine differentiation, sufficient market potential, a sales team with expertise, and a management team willing to endure bad target accounts. Without these prerequisites, it becomes difficult.

It also doesn't prove that inbound is dead. But anyone who, as a technical niche provider, only waits to be found leaves the market to chance. Especially companies that not everyone knows need proactive market cultivation. Not aggressive. Precise.

And it doesn't prove that AI replaces technical consulting. On the contrary. The better AI does the preparatory work, the more important human consulting becomes at the right moment. This is the shift I see in good implementations: less human time at the beginning, more human quality where purchasing decisions are really made.

My Forecast for Manufacturing SMEs

I believe that the technical SME sector will experience a tough separation in the next two years. Companies that systematize their market and ICP knowledge will generate more pipeline with smaller teams. Companies that continue to store their knowledge in heads and Excel files will talk more about weak demand. Although the market is not the main problem.

That sounds strict. And it should. SMEs have enough real problems: energy costs, regulation, raw material risks, skilled labor shortages, investment reluctance. But precisely for this reason, sales must not be treated as a secondary process. Pipeline is not a gift from the market. Pipeline is an operating system.

Magnet World is a good example for me because the story is not exaggerated. Not a huge corporation. No software product with viral growth. No marketing team with 30 people. Just a specialized provider with real technical substance, a narrow niche, and the will to make its market cultivation cleaner.

In the end, one sentence from the review stuck with me. Tobias said: "We didn't sell more because we talked more. We talked to the right people earlier." You can hardly describe AI in sales for SMEs better than that.

Full Success Story The complete Magnet World success story with background, implementation, and results of the Amplifa implementation.