AI Sales: Schnaithmann Scales Lookalikes

Case Study · 16. Mai 2026 · Mohsen Ghulami

AI Sales in mechanical engineering: Read how Schnaithmann finds 300 lookalikes per week, doubles demos, and cleanly compares providers. Practical insights.

Last Thursday, 7:42 AM, I'm standing in a Schnaithmann hall in Remshalden. Next to me is Tobias, Head of Sales. The coffee smells like it's from a thermos, a pneumatic cylinder is tapping somewhere every second, and on his screen is a list of companies that I used to painstakingly copy from trade fair catalogs. "300 lookalikes per week," he says, tapping his finger on the column "similar production environment." I ask for clarification because I rarely let such statements go unchecked in mechanical engineering halls. AI Sales here is not just a buzzword; it's a rather sober answer to an old problem: How does a special machine builder find the next good factory before purchasing has invited three competitors?

Schnaithmann Maschinenbau GmbH, a special machine builder and automation partner from Southern Germany, is not a startup with beanbags nor a corporation with 40 people in marketing. Several hundred employees, transfer and conveyor systems, assembly automation, special machines, control technology, projects in automotive, electronics, and general industry. In other words, precisely the kind of medium-sized company I've been visiting since 1998: technically strong, often dependent on existing customers, trade fairs, and the famous "I know someone at Brose" for sales. That works. Until it's no longer enough.

Why this AI Sales Comparison is Necessary in Special Machine Building

In March 2025, I sat with Andrea, Head of Sales at an automation company in Bielefeld, in a hotel by the A2. The breakfast buffet clattered beside us, she pulled up her CRM and said: "Our best projects start six months before anyone writes an inquiry." That's precisely the point. In special machine building, demand doesn't arrive like a package with a tracking number. It appears as a job advertisement for Industrial Engineering, as a new hall in Hungary, as a product change at a Tier-1 supplier, as a capacity problem in an assembly cell that is suddenly no longer shown during a factory tour.

The old logic was: trade fair, network, inquiry, offer, follow-up. The new logic is more uncomfortable. Anyone who still believes in 2026 that inbound alone will secure the pipeline confuses awareness with market coverage. According to the VDMA business survey for mechanical engineering from Q2 2025, 58 percent of surveyed companies cited a weak or uncertain order intake situation as a central risk, while skilled labor shortages and automation pressure continued simultaneously. There's no getting around it: the need is there, but it's not neatly sitting in a contact form.

Schnaithmann's starting point was therefore typical. Good references. Strong technology. A sales team capable of conducting technical discussions, but not identifying hundreds of potentially suitable factories in DACH and Europe every week. Tobias described the pre-state without drama: "We didn't have a lead crisis. We had a coverage crisis." That's a difference. A lead crisis sounds like a marketing problem. A coverage crisis means: the market is moving in places you're not currently looking.

Evaluation Criteria: How I Measure AI Sales in Mechanical Engineering

Over the past few years, I've seen dozens of sales software demos. Some looked like a space station and solved the problem of a simple Excel export. Others were ugly but effective. For Schnaithmann and comparable companies like Wittenstein, Festo-affiliated integrators, or smaller automation companies around Stuttgart, it's not about whether a platform draws pretty dashboards. What matters is whether sales can get into real projects earlier.

For this comparison, I used seven criteria:

• ICP Sharpness: Does the system really identify suitable factories, lines, components, and roles – or just companies with a matching NACE code?
• Lookalike Search: Can the solution derive new similar target accounts from existing good customers, such as automotive suppliers with assembly and testing processes?
• Data Quality in the Buying Center: Does it find production managers, industrial engineers, plant managers, and project managers – not just managing directors and marketing contacts?
• Personalization in Outreach: Does the approach sound like mechanical engineering and a specific process problem, or like generic software sales?
• Workflow Depth: Are research, prioritization, sequences, follow-ups, and CRM handover integrated?
• Measurability: Are there hard numbers for appointments, demos, concept workshops, response rates, and pipeline – not just open rates?
• Suitability for SMEs: Does the solution fit a sales team of five to 25 people, a limited marketing department, and long sales cycles?

These criteria sound dry. In practice, they decide months. At a machine builder from the Backnang area, Jens, the managing director, told me in January 2025: "If we show up two months too late for a battery module project, we're politely invited, but technically already out." He said that during a factory tour, while a stacker crane squeaked behind us. You remember such sentences.

Candidate 1: Amplifa as an AI Sales Engine for Schnaithmann

What Amplifa Specifically Did at Schnaithmann

Amplifa did not appear at Schnaithmann as another newsletter tool. That would be ridiculous in special machine building. The task was more precise: derive target patterns from existing successful projects, automatically find lookalike accounts based on these patterns, and approach them with multi-stage, personalized outreach. Fully automated lookalike search and outreach – sounds like a pitch. At Schnaithmann, it became a weekly routine.

The basis was existing customers and project types: assembly and handling automation, transfer technology, line linking, testing cells, applications for automotive suppliers, electromechanics, and industrial assemblies. Amplifa didn't just pull a crude industry list from this, but search logics: Which factories have similar manufacturing processes? Which companies build comparable components? Where are there indications of capacity expansion, new product lines, or job profiles for automation? According to the Schnaithmann project context, this resulted in around 300 new lookalike target accounts per week. Not 300 "leads" in the confetti sense. 300 verifiable companies or locations that were similar enough to warrant a genuine initial approach.

Tobias showed me a before-and-after view in the CRM. Before Amplifa, many new conversations came from trade fairs like Motek Stuttgart, existing customer recommendations, and occasional inbound inquiries. After implementation, new accounts flowed weekly into prioritized sequences, including contacts in production, engineering, and plant management. Within six months, Schnaithmann doubled the number of qualified demos and concept discussions. That's the number that sticks. Not because it's magical, but because in a market with a 6 to 24-month sales cycle, it takes effect early enough to become order intake later.

We no longer wanted to guess which plants suited us. We wanted to see every week which accounts looked like our best customers – and then approach them cleanly, technically, and respectfully.

— Tobias, Head of Sales at Schnaithmann, Remshalden

Strengths of Amplifa in Mechanical Engineering Sales

Amplifa's strongest side, in my opinion, lies in the combination of market intelligence and execution. Many tools find names. Others send emails. This sounds like a division of labor, but in SMEs, it often ends up as semi-finished goods: data in one system, sequences in another, CRM half-maintained, sales annoyed. Amplifa tries to close the chain – target account, buying center, occasion, approach, appointment handover.

At Schnaithmann, it was particularly important that the approach didn't sound like a SaaS company. A production manager at a Tier-1 supplier in Zwickau doesn't want to read an email saying they can "unlock growth potential." They want to know if someone understands their problem: cycle time, variant changes, scrap, operator shortages, linking between stations. Amplifa used modular text blocks, industry hints, and account-related triggers for this. Not every email will be a poem. It doesn't have to be. But it shouldn't reek of a mass mailing.

The weakness? Amplifa requires clear thinking at the beginning. Anyone who cannot define their Ideal Customer Profile, who finds all industries somehow interesting, and who only tells the system "more leads please" will scale nonsense. Well, almost. Amplifa at least forces such companies to make this imprecision visible. But a sales manager must decide: Which projects do we really want? What margin do we need? Which use cases fit the engineering capacity? AI does not remove this leadership task.

Amplifa Product AI sales platform for lookalike search, buying center research, personalized outreach, and CRM-integrated sales workflows.

Candidate 2: Classic Sales Stack of CRM, Database, and Sequencing

HubSpot, Apollo, Cognism, Salesloft – strong, but often too generic

The second candidate is not a single tool, but the classic setup I see at many mechanical engineering companies: HubSpot or Salesforce as CRM, plus a data provider like Cognism, Dealfront, or Apollo, plus perhaps Salesloft, Lemlist, or Outreach.io for sequences. At a packaging machine manufacturer near Schwäbisch Hall, Martin, CSO, showed me four browser tabs in April 2025 and said: "We've actually bought everything. But nobody's using it." An old laser printer hummed next to his desk. A very German picture.

These stacks can work well if a company has a clean Revenue Operations team. Cognism provides European contact data, HubSpot maps campaigns and pipeline, Salesloft controls sequences, LinkedIn adds context. For larger companies, such as sales organizations affiliated with Schaeffler or Phoenix Contact with clear roles, this is solid. You can configure, test, segment, and measure a lot.

In special machine building, that's precisely the problem. Configuration is not a result. If three sales engineers are supposed to clean lists, find contacts, write emails, and maintain the CRM after customer appointments, it won't happen. Or only in September, when panic sets in after the summer break. In addition, databases rarely understand manufacturing similarity. They find "Automotive Supplier" or "Machinery," but not necessarily a plant with manual pre-assembly, a test cell, high variant diversity, and an upcoming line for e-mobility components.

The strength of the classic stack is control. The weakness is friction. You get building blocks, not a machine. For companies with internal Sales Ops, this can be the right approach. For Schnaithmann, in my estimation, it would have been putting the cart before the horse: first build the tool landscape, then hope that new qualified conversations arise weekly. I've seen this hope wither in CRM too often.

We had 18,000 contacts in the system and still too few new initial conversations. That's not a data problem, that's a process problem.

— Martin, CSO of a packaging machine manufacturer, Schwäbisch Hall

Candidate 3: LinkedIn Sales Navigator for Manual Market Development

LinkedIn Sales Navigator has now almost arrived everywhere in German mechanical engineering. I see the tool at Festo suppliers, robotics integrators in Augsburg, and special machine builders in Sauerland. It's good for making people, job changes, company updates, and networks visible. Sarah, Business Development Manager in Nuremberg, told me in February 2025: "Without LinkedIn, I find some production managers faster than through the switchboard." True. But Sales Navigator remains a search instrument. It doesn't automatically think in lookalike factories, it doesn't build a reliable account logic from Schnaithmann's best projects, and it doesn't relieve anyone of the weekly research block.

For very experienced salespeople, LinkedIn is a scalpel. For understaffed teams, it quickly becomes occupational therapy. You scroll, save leads, write three messages, get a reply with "gladly after Motek," and lose the thread. I'm not against LinkedIn. On the contrary. But as a sole solution for AI Sales in mechanical engineering, it's too narrow. It shows people, not necessarily project windows.

Candidate 4: Trade Fairs, Inbound, and Lead Agency – the Old Machines

Trade fairs like Motek, automatica, or LogiMAT remain important. Anyone who claims otherwise hasn't been in a hall for a long time where a customer with oily fingers is fiddling with a transfer system and asking if the workpiece carrier also works with their component geometry. Schnaithmann, like many automation companies, benefits from such moments. But trade fairs are pacemakers, not market coverage. A Motek week in Stuttgart doesn't replace 50 weeks of systematic account scouting.

Lead agencies, in turn, can deliver appointments in the short term. Some good, some bad. In June 2025, I saw the evaluation of a machine builder from Upper Franconia: 1,200 cold emails, four replies, one of which was a complaint about the wrong industry. The managing director, Thomas, put the printout in front of me and said: "We're stopping that now." Understandable. But outbound itself is not wrong. Bad outbound without mechanical engineering context is wrong.

Large Comparison Table: AI Sales, Sales Stack, LinkedIn, and Trade Fairs

The following table is deliberately not academic. It is based on conversations with sales managers from manufacturing SMEs, on my observations at Schnaithmann, and on typical cost and process structures as I see them between Stuttgart, Bielefeld, and Nuremberg. Status: July 2025. Anyone looking for absolute truth here will be disappointed. Anyone who needs a decision-making basis will get one.

Price Comparison: What Does Pipeline Really Cost?

In B2B sales, prices are often discussed as if they were about screws. Wrong benchmark. The relevant question is not: What does the software cost per month? But rather: What does a qualified conversation with a plant that could automate in the next 6 to 18 months cost? For special machine builders with project volumes from 250,000 euros to several million, a single early access is worth more than half a year of tool discussion.

Nevertheless, a managing director needs numbers. So roughly, as of summer 2025, typical market observations from DACH, without individual discounts and implementation details. I'm giving ranges because exact prices vary depending on the number of users, data package, and service component. Anyone who claims there's a simple list price comparison here has never had coffee with enterprise SaaS salespeople.

I've seen CFOs flinch at "AI investment" and in the next breath approve 120,000 euros for a trade fair stand because the stand has been there for 14 years. Not entirely true, some are asking tougher questions now. But the cultural asymmetry remains. Well-known expenses are considered solid, new expenses have to lay themselves bare on the table. That's human, but dangerous.

What Didn't Work Before at Schnaithmann

Schnaithmann didn't have sales chaos. That's important to me. It wasn't the story of a company buying AI out of desperation. Rather, it was the story of a technically strong special machine builder who realized that his traditional search pattern was becoming too small. Automotive is changing, e-mobility is shifting value creation, electronics manufacturers are automating, Eastern European plants are growing, German plants are modernizing under pressure from skilled labor shortages.

Before Amplifa, market development was heavily focused on known clusters. Existing customers, recommendations, trade fair contacts, occasional RFQs. Plus individual manual research when a sales engineer found time. But a sales engineer rarely finds time because they're stuck in layout discussions, checking cycle time calculations, or standing in front of a line at a customer's site that isn't running as described in the specifications. Anyone who sells cleanly in this profession doesn't have half-hearted technical conversations.

Tobias put it bluntly: "Our best people were too expensive for Google research." Exactly. And too good. I've seen this repeatedly at DMG Mori suppliers, at small robotics cell builders in Bavaria, and at testing machine manufacturers in Baden-Württemberg: the most expensive sales resource spends too much time on the beginning of the search instead of the end of qualification. An experienced sales engineer's hour belongs in a conversation about process, feasibility, budget, and timing. Not in the question of whether company X in the Czech Republic might have a new assembly hall.

How the Amplifa Workflow Ran at Schnaithmann

From Existing Customers to 300 Lookalikes per Week

The first step was not AI magic, but homework. Schnaithmann and Amplifa defined which existing customers and projects served as blueprints. Not every revenue is good revenue. Some projects are technically appealing and commercially tough. Some customers fit the capacity, others drag engineering into unpaid concept loops for weeks. So patterns were worked out: industries, applications, plant sizes, roles, typical triggers, regional focuses.

After that, Amplifa searched for similar companies and locations. For mechanical engineering, this means: not just company profiles, but clues from websites, job advertisements, news, production descriptions, trade fair exhibitor lists, supply chains, and role patterns. If a company is looking for new industrial engineers for assembly automation, that's not an order. But it's smoke. If a new product line is added and the company already describes manual assembly processes, it becomes more interesting.

The approximately 300 lookalikes per week were not blindly thrown into sequences. Good systems prioritize. Bad systems blast. At Schnaithmann, accounts were sorted by fit, presumed occasion, and contact quality. Then outreach sequences ran, mostly via email and LinkedIn support, in German and for export markets in English. The goal was not immediate machine sales. The goal was a qualified initial conversation, a demo, a concept workshop, an early seat at the table.

1. Select reference projects: Which customers, applications, and machines were technically and commercially truly attractive?
2. Translate ICP: Gut feeling became criteria like industry, plant type, manufacturing process, roles, and triggers.
3. Search for lookalikes: Amplifa weekly identified similar accounts and locations in DACH and adjacent markets.
4. Find Buying Center: Production management, engineering, plant management, and project managers were prioritized.
5. Personalize outreach: The approach referred to the manufacturing environment, automation pressure, or specific signals.
6. Measure handover: Appointments, demos, and concept discussions were entered into the CRM and checked for pipeline contribution.

AI Sales in Sales: What the Demo Doubling Really Means

"Demo doubling in six months" sounds like a standard slide in software sales. In special machine building, it's different. A demo here is rarely a 20-minute webinar with screen sharing. It can be a technical exchange, a remote concept appointment, a workshop on cycle times, or an initial discussion about line linking. If these conversations double, order intake doesn't automatically increase in the same month. But the probability of getting into the planning process earlier increases.

And early planning processes are the real currency. For automotive suppliers, special machines are often discussed long before a formal tender. Production, industrial engineering, and purchasing sort providers before a specification is finalized. Those who have already delivered technical ideas then carry weight. Those who only appear at the RFQ often only get to play price anchor. That's tough, but true.

I asked Tobias if the additional conversations were also qualitatively appropriate. He didn't show me a celebratory graphic, but three examples: an electromechanics manufacturer with manual test assembly, a Tier-2 supplier with a variant problem, a plant in Eastern Europe with planned capacity expansion. None of them were orders yet. But all were conversations that Schnaithmann probably wouldn't have had at this stage without systematic lookalike search. That's exactly what it's about.

FAQ: Does AI Sales Even Work for Complex Machinery?

Yes, if you don't confuse AI Sales with automatic selling. An AI doesn't sell a special machine for 1.4 million euros to a plant manager in Ingolstadt. Ridiculous. But it can find accounts, sort signals, research contacts, prepare initial approaches, and execute follow-ups diligently. Technical sales remain human. But the human enters the process later – where they create value.

FAQ: Why are Trade Fair Contacts No Longer Enough?

Because trade fair contacts are sporadic. At automatica 2024 in Munich, Frank, sales manager of a gripper manufacturer from the Heilbronn area, told me between two hall sandwiches: "We see many right people here, but not all the right times." You can't say it better. Trade fair is density, not timing. AI Sales tries to improve timing through signals and permanent search.

FAQ: What is the Biggest Mistake in Lookalike Lead Generation?

The biggest mistake is defining similarity too cheaply. "Automotive" is not similarity. "More than 500 employees" isn't either. It only becomes interesting with process proximity: assembly of electromechanical components, high variant mix, manual test stations, new line, operator bottleneck, existing conveyor technology, investment window. Anyone who only uses industry filters produces lists. Anyone who uses process similarity produces opportunities.

Three Learnings from the Schnaithmann Case Study

Learning 1: The Best ICP is in the Best Projects

Many sales managers start ICP work with desired customers. A nice mistake. It's better to look at completed projects: Where were the margin, collaboration, technical fit, and follow-up potential right? At Schnaithmann, existing projects were the starting point for the lookalike logic. This sounds trivial, but it is rarely consistently implemented. I have seen CRM systems where project reasons, applications, and decision-making roles could not be evaluated at all. Only revenue, date, contact person. For AI Sales, that's too thin.

Learning 2: Sales Engineers Don't Belong in Cold Research

A good sales engineer can find out in 45 minutes whether an automation project has substance. He hears from a side comment whether a budget exists. He notices whether production and purchasing are working against each other. He recognizes whether the cycle time requirement is physically nonsense. Precisely this ability should not be wasted on address research. Amplifa did not take over sales from Schnaithmann, but the preliminary work. That's a difference that managing directors should understand.

Learning 3: Outbound Must Sound Technical, Not Loud

Mechanical engineering doesn't hate bad sales; it hates superficial sales. A cold approach can be cold if it's respectful and relevant. "We build assembly automation for companies with variant and cycle time pressure in electromechanical assemblies" is better than "We help you realize efficiency potentials." The latter can go. The former can be checked by a production manager.

Personal Recommendation: Which Solution I Would Choose

For a special machine builder like Schnaithmann, I would choose Amplifa over a classic self-service stack. Not because HubSpot, Cognism, or Salesloft are bad. They're not. But they require an organizational maturity that many SMEs in sales don't yet have – and perhaps don't need. If the goal is to find new suitable plants weekly, identify buying centers, and generate qualified conversations, execution matters more than tool ownership.

My second recommendation is more uncomfortable: Anyone implementing Amplifa or a similar AI sales system must first decide which business they no longer want. Otherwise, the funnel will get fuller, but not better. Schnaithmann's advantage lay not only in the technology but in the willingness to derive hard patterns from successful projects. That's sales strategy. No tool can fake it.

Amplifa Sales Audit Analysis for manufacturing SMEs: Where does your sales lose market coverage, timing, and qualified initial conversations?

Decision Aid: Three Questions Before Tool Selection

Before any investment, I would put three questions on the table. Not in a workshop with 19 sticky notes. In a room with sales, management, and someone from engineering who really knows the market. At Schnaithmann, this mix was clearly at the table. You can tell from the answers.

1. Can we describe our ten best customers and projects in such a way that a system can find similar plants from them?
2. Do our sales engineers currently have enough qualified initial conversations – or do they spend too much time searching, listing, and following up?
3. Do we want more volume at any cost, or do we want to get into the right investment windows earlier with suitable accounts?

If the answer to question one is "no," the work begins in the CRM and project analysis. If question two hurts, that's a good sign. Pain often only shows where the process becomes honest. And if question three is not cleanly answered, every lead generation becomes a lottery with a pretty interface.

The Business Effect: More Market Coverage Without More Salespeople

The word scaling is often misused in sales. At Schnaithmann, it doesn't mean that robots suddenly sell. It means: Sales sees more suitable market movements, approaches more relevant accounts, and conducts more early conversations, without immediately hiring new sales staff. In a labor market where good technical salespeople are scarce, this is not a side effect. This is the core.

According to the Federal Employment Agency, the bottleneck in technical professions in 2024 was significantly higher than in many commercial profiles; mechanical engineering companies have been reporting difficulties in finding experienced sales engineers for years. In May 2025, I spoke with Peter, managing director of an automation company in Regensburg. He said: "I can buy a milling machine more easily than find a good man for technical sales." The sentence was uttered next to a box of Siemens S7 modules. It wasn't meant to be funny.

That's why AI Sales is so interesting for SMEs. Not as a replacement for people, but as leverage for scarce people. Schnaithmann's demo doubling in six months shows how top-of-funnel work can be industrialized without devaluing technical sales. A production manager ultimately wants to talk to someone who understands their line. But that someone first needs to be invited.

Where Amplifa Doesn't Fit

I don't trust any solution that supposedly always fits. Amplifa is not ideal for companies that serve extremely broad mass markets and only want to maximize simple contact forms. Nor for companies that cannot explain their own offering. If sales internally doesn't know whether to prioritize components, projects, retrofits, or consulting, AI will accelerate this indecision. Faster wrong is still wrong.

Very early companies without reference projects also struggle more, because lookalike search thrives on patterns. Without patterns, only guesswork. For established SMEs like Schnaithmann, that's precisely the advantage: they sit on 20, 50, or 100 projects from which market logic can be derived. Only this logic was previously often stored in the minds of two senior salespeople. One retires in 2027. Then what?

Why Schnaithmann is a Good Example for SMEs

Schnaithmann is neither the largest nor the loudest automation company in the country. That's precisely why the story is interesting. A special machine builder from Southern Germany, rooted in mechanics, control, project business, with customers in automotive and industry, professionalizes its go-to-market without losing its identity. No drama. No AI show. Rather, a new machine in sales that runs every week.

As I left, I thought of a scene from the hall. An employee pushed a workpiece carrier over a transfer system, the plastic quietly scraping on the guide, and Tobias looked at the account list again. There were companies from Baden-Württemberg, the Czech Republic, Austria, North Rhine-Westphalia. Not all will become customers. Some will never respond. But a few will have a problem at precisely the moment when Schnaithmann would not have been visible before. In special machine building, sometimes that exact moment is enough.

Full Success Story Read the complete Schnaithmann case study: 300 lookalikes per week and demo doubling in six months.