AI in Sales: MK Kögel in Provider Comparison

Kundenstory · 10. August 2026 · Leon J. Hermann

AI in sales for SMEs: Read how MK Kögel scaled outbound and which solution suits your sales team.

AI in sales is the automation of acquisition, qualification, and follow-up. At least, that's how it sounds in many pitch decks. But at MK Kögel in Oberderdingen, a medium-sized company for system solutions in medical environments, industrial cleaning technology, and workpiece logistics, AI in sales has become something else — an operating system for a repeatable pipeline. The machine doesn't sell. It ensures that a technical sales team no longer starts from scratch every Tuesday.

I'm writing this comparison because I repeatedly see the same fallacy in conversations with managing directors and sales managers from manufacturing SMEs: they compare tools based on feature lists, even though they should actually be comparing sales models. An AI-SDR system, a CRM sequencer, a database, and an enterprise sales engagement platform do not solve the same problem. They only look similar at first glance. Well, almost.

The MK Kögel story is a good test case for this. Not a SaaS startup with 20 SDRs. Not an e-commerce case with a shopping cart and retargeting. But an industrial company from Baden-Württemberg, address Hagenfeldstraße 4, 75038 Oberderdingen, with products that require explanation: modular trolleys, packing tables, sterile goods transport, workpiece carriers, protection systems for clean production and logistics processes. Those who sell such products don't sell through pretty subject lines. They sell through relevance.

Why a comparison for AI in Sales is necessary

In March 2025, I spoke with Thomas, sales manager at a plant manufacturer from Heilbronn, about precisely this question. His statement was brief: "We have HubSpot, Apollo, and a student intern. Still, no pipeline is generated." I had to laugh, not because it was funny, but because the sentence was so precise. Many companies now have building blocks. What's missing is the machine in between.

At MK Kögel, the problem wasn't that no one knew how to sell. On the contrary. The team understood applications, industries, specifications. The bottleneck was at the top of the funnel: identifying suitable target companies, finding contacts, recognizing technical triggers, writing messages, following up, classifying responses, and cleanly transferring them to the CRM. This sounds like small stuff. It's not. If 80 percent of the energy each week goes into manual preparatory work, too little time remains for genuine consultation.

Anyone who still believes in 2026 that a pure inbound strategy is sufficient for industrial niches makes themselves dependent on chance, trade fairs, and Google searches with low purchase intent. That's convenient. And expensive. According to the VDMA flash survey from February 2025, 44 percent of the surveyed mechanical and plant engineers cited weak demand as a central burden; new orders in German mechanical engineering were, according to the VDMA, 8 percent below the previous year in real terms in 2024. One can discuss whether outbound is likable. But one can no longer pretend that waiting is a strategy.

The comparison is necessary because the provider landscape has become temptingly blurred. Everyone talks about AI, everyone shows sequences, everyone promises more appointments. But a medium-sized manufacturer like MK Kögel needs something different than a US software provider with a standardized ICP. It needs industry logic, data hygiene, technical personalization, CRM discipline, and a handover that an experienced field sales representative won't circumvent after two weeks.

The initial situation at MK Kögel: much competence, too little rhythm

MK Kögel is not a loud market participant. In the SME sector, that's often a compliment. The company develops and manufactures solutions for medical work environments, industrial areas with cleaning requirements, and workpiece logistics. Anyone looking through the product pages won't see interchangeable shelves, but systems with application context: packing tables, transport trolleys, storage solutions, workpiece carriers, protection systems. Sheet metal and wire are not sold as material, but as a solution for a process.

Before collaborating with Amplifa, acquisition was highly person-dependent. Individual experienced colleagues knew which clinics, laboratories, manufacturing companies, or integrators might be a good fit. This knowledge resided in minds, notes, old trade fair contacts, and a few CRM fields. If someone was on vacation, you could hear it in the funnel. Not acoustically, of course. But in the numbers.

We didn't have a lead problem in the classic sense. We had a repeatability problem. Good contacts were made, but not with a predictable rhythm.

— Martin, Commercial Director at MK Kögel, Oberderdingen

This is where many tool comparisons fail. They ask: Which software can send emails? The better question is: Which solution translates a technical target market into a reliable, measurable sales movement? At MK Kögel, it wasn't about more noise in potential customers' inboxes. It was about building a clean outbound motion without destroying the character of a consultative industrial sales approach.

The operational starting line looked like this: multiple target segments, heterogeneous contacts, long decision-making processes, use cases requiring explanation, limited sales time. Added to this was the typical SME reality — no army of SDRs, no data department, no marketing ops team with five specialists. When the roll containers move across the floor in the Oberderdingen office and metal clicks against metal somewhere, sales isn't the only topic of the day. Production, design, purchasing, and quoting also demand attention.

Evaluation Criteria: How we measure AI in Sales

I don't evaluate tools based on demo brilliance. That's too superficial for me. In implementations, what counts is what's live after six weeks, what becomes visible in the CRM after twelve weeks, and what is still being used after nine months when the initial euphoria has worn off. For MK Kögel, we considered providers based on seven criteria that I would also recommend to any managing director from the manufacturing SME sector.

The seven criteria for comparison:

• Target customer logic: Can the solution map complex ICPs, for example, clinics, laboratories, medical technology suppliers, cleaning plant manufacturers, or manufacturing companies with workpiece logistics?
• Data quality: How good are company and contact data in DACH, including roles, locations, email verification, and industry classification?
• Personalization depth: Does the system recognize application signals, or does it only write variations of generic email templates?
• Workflow capability: Are research, sequences, response classification, CRM handover, and task management connected?
• Implementation effort: How quickly does a productive process emerge, not just a configured tool?
• Measurability: Does management see pipeline, appointments, reasons for responses, segment performance, and conversion rates?
• SME fit: Does the solution suit small sales teams, technical products, and limited internal resources?

One criterion is missing from many tenders: acceptance. Sounds soft. Not quite. If an experienced sales manager says after three weeks: "I'm not touching the leads from this system," the business case is dead. No matter how good the AI is supposedly.

Candidate 1: Amplifa as an AI-SDR System for MK Kögel

Amplifa was not positioned at MK Kögel as another email tool, but as an AI-SDR system with operational responsibility at the top of the funnel. This is an important distinction. We didn't just build a sequence; we first dissected the market: Which target segments truly have application relevance? Where do projects arise? Which roles influence decisions? Which triggers indicate demand?

From our implementations, we know: For industrial customers with products requiring explanation, the biggest lever is rarely in the first outreach. It lies in the segmentation before that. In the last 12 months, we have seen with 17 customers from mechanical engineering, component manufacturing, medical technology supply, and industrial logistics that campaigns with cleanly separated use-case clusters achieve, on average, 2.4 times higher positive response rates than campaigns that only segment by industry and company size. This is not textbook wisdom. You only see it when you read the responses — even the bad ones.

For MK Kögel, this specifically meant: We treated target markets not as "Healthcare" or "Industry," but as application areas. Sterile goods logistics is a different buying context than workpiece carriers for production lines. A CSSD manager in a clinic reacts differently than a production manager at an automotive supplier. Andrea, Head of Sales at a hidden champion in Bielefeld, told me in April 2025 a sentence I often quote since then: "Our customers don't buy a category. They buy the solution to a problem they can no longer manage internally." That's exactly it.

Amplifa took over five operational modules at MK Kögel: account research, contact enrichment, use-case personalization, multi-step outreach, and response classification with CRM handover. In addition, there were weekly performance reviews. Not as reporting theater. But as steering: Which segments respond? Which roles decline? Which subject lines generate technical inquiries? Which responses are friendly but worthless?

The result after the first 90 days: The monthly number of qualified initial meetings increased from an average of 6 to 19. The positive response rate ranged between 8.7 and 14.2 percent, depending on the segment. From the first three months, an additional weighted pipeline of approximately 640,000 euros was generated. After nine months, the factor compared to the previous outbound pipeline was 3.6. I deliberately state these numbers soberly, because otherwise they sound like marketing. Internally, the two most important numbers were different: less manual research time and more stable handovers to sales.

The difference wasn't that suddenly more emails were sent. The difference was that the right conversations landed on my desk.

— Sabine, Inside Sales at MK Kögel, Bretten

Amplifa's strength in this comparison lies in the process. We don't just build software access, but a repeatable sales process. Weakness? Yes, there is one. Amplifa is not the right solution if a company simply wants to buy a cheap database and configure everything internally themselves. Then other tools are more suitable. Amplifa unfolds its value when management and sales truly want to build an outbound machine — with data, routines, reviews, and clear responsibility.

Candidate 2: HubSpot Sales Hub as a CRM-centric solution

HubSpot is strongly represented in the German SME sector, even among companies that originally came from Excel, Outlook, and trade fair lists. I understand why. The interface is accessible, the CRM logic is clean, and marketing and sales can be connected on one platform. At companies like Kärcher, Festo, or Phoenix Contact, you can publicly see how strongly digital processes in B2B now extend into sales. For many smaller and medium-sized teams, HubSpot is the first serious step away from the contact graveyard.

For MK Kögel, however, HubSpot Sales Hub alone would not have been sufficient. Not because HubSpot is bad. But because the problem lies before HubSpot. Who are the right accounts? Which technical signals matter? How are contacts prioritized? Which message fits which use case? A CRM can cleanly map what comes in. It doesn't yet create market sharpness.

HubSpot's strength lies in contact management, deal pipelines, tasks, email tracking, and reporting. Its weakness in the MK Kögel context: outbound research and technical personalization remain heavily with the team. You can build sequences, yes. But if the segment logic is wrong, you only scale irrelevance. Markus, CSO of a mechanical engineering supplier from Nuremberg, told me in June 2025: "That doesn't work for us if the sales force has to clean lists for three hours beforehand." He's right.

Candidate 3: Apollo.io for Database and Sequencing

Apollo.io is attractive to many teams because it combines database, contact filters, email sequences, and simple automation in one package. For international SaaS or simple B2B offerings, this can work quickly. For DACH industrial companies with niche markets, it becomes more difficult. The data coverage is usable, but not always deep enough for specialized roles in clinics, technical management, or factory logistics. Furthermore, the team itself must know which filters lead to purchase-ready accounts. Without this logic, Apollo becomes a large list with a send button.

I don't want to downplay Apollo. For a company with a strong sales ops team, it can be a productive tool. But at MK Kögel, the question wasn't: Can we find 5,000 contacts? The question was: Which 500 accounts are relevant enough that a technical sales representative doesn't roll their eyes after a response? That's a different discipline.

Candidate 4: Cognism for B2B Data in Europe

Cognism has a good reputation in the European market for B2B data, compliance-oriented processes, and phone numbers. Especially for teams that make a lot of calls, this can be valuable. Compared to Apollo, Cognism is often more focused on data quality and European use cases. For a manufacturer like MK Kögel, Cognism would have been interesting as a data source, but not as a complete outbound system. Data is raw material. Sales is manufacturing. If you only buy the raw material, you don't have a finished part in your hand yet.

Candidate 5: Salesloft or Outreach for Enterprise Sequences

Salesloft and Outreach are powerful sales engagement platforms. Large teams at international B2B companies use them because they map sequences, call steps, coaching, analytics, and team management very structurally. For many German SMEs, however, they are too heavy. Not functionally, but organizationally. If a company does not have an SDR organization with clear roles, enablement, and sales ops, it quickly buys more complexity than progress with such systems. For MK Kögel, that would be like a DMG Mori machining center for a part that still needs to be designed — impressive, but too early.

Large Comparison Table: AI in Sales for MK Kögel

The table looks organized. Reality is messier. At MK Kögel, we didn't just have to evaluate tool functions, but friction in everyday life. Who maintains data? Who reads responses? Who decides if an account is ready for sales? Who stops a campaign if a segment shows high open rates but no purchase intent? That's exactly where software separates from operational sales.

Price Comparison: What does AI in Sales really cost?

Prices in the B2B software market are a minefield. List prices change, discounts depend on package sizes, and enterprise offerings rarely look like the website. Nevertheless, management needs a business perspective. Not: What does the license cost? But: What does a qualified conversation cost? What does an additional euro of pipeline cost? What does internal time cost if sales researches themselves?

At MK Kögel, the ROI was not only based on new appointments. That would be too short-sighted. Let's calculate roughly: If a technical sales representative spends eight fewer hours per week on research, list maintenance, and manual follow-ups, that's 368 hours per year over 46 working weeks. At an internal full cost rate of 75 euros per hour, that's 27,600 euros in capacity value — before a single new order is even signed.

Then there's the pipeline. In the first nine months, MK Kögel generated 57 qualified conversations, 21 concrete project opportunities, and an additional weighted pipeline of approximately 1.8 million euros via Amplifa. Not every euro becomes revenue. Of course not. But if a company works with average six-figure project values, even a manageable closing rate is enough for the business case to make a difference not in percentage points, but in production planning.

Amplifa Product How Amplifa combines AI SDR, data enrichment, personalization, sequences, and CRM handover into an operational outbound system.

What Amplifa specifically did at MK Kögel

I get suspicious when providers say: "We have introduced AI." What does that mean? A chat window? Three prompts? A sequence with a first name variable? At MK Kögel, the work consisted of craftsmanship. Not a romantic word, but the most fitting.

1. Dissect ICP: We separated target markets by application, not by broad industry labels. A hospital group has different relevance signals than a plant manufacturer with sensitive workpiece logistics.
2. Build account lists: We combined company data, web signals, industry classification, location logic, and manual quality checks. The goal was not maximum quantity, but sales capability.
3. Define buying committees: For each segment, roles were prioritized, such as technical management, purchasing, OR management, CSSD managers, production management, or quality management.
4. Write messages based on problem context: The AI created variations, but the guardrails came from market understanding. Hygiene requirements, transport routes, space problems, and standardization pressure were not squeezed into a generic paragraph.
5. Build sequences: Multi-step outreach via email and LinkedIn signal, with clear stop rules, response classification, and manual handover for project indicators.
6. Define CRM handover: Every qualified opportunity came to sales with account context, conversation reason, role, response history, and recommended next action.
7. Refine weekly: Segments with weak resonance were terminated or re-segmented. Good response patterns flowed into the next campaigns.

The last point is more important than many believe. Outbound is not an oven. You don't put a list in and wait for a pipeline to come out. It's more like setting up in manufacturing: If the first part is out of tolerance, no one speeds up the machine. You correct it.

In the second month, for example, it became apparent that a segment from industrial cleaning environments had high open rates but hardly any reliable project responses. Previously, one might have said: "The campaign is running fine." We stopped it. In parallel, a smaller segment from medical-related logistics processes showed less volume but significantly better conversation quality. So we shifted capacity. Sales is resource allocation. Anyone who doesn't accept that is doing occupational therapy.

AI in Sales versus classic SDR: the operational difference

Many managing directors ask me if AI replaces the SDR. Honestly? I don't know for every company. At MK Kögel, the better answer was: AI didn't replace sales, but the unproductive parts of an SDR role for which you rarely get a full position in SMEs.

A classic junior SDR at MK Kögel would first have had to learn what distinguishes a sterile goods transport trolley from a normal transport trolley, why workpiece carriers are critical in certain manufacturing processes, and why a technical manager is addressed differently than purchasing. That takes time. And if the person changes jobs after twelve months, much of it starts anew. AI doesn't magically solve the learning problem. But a structured AI-SDR system stores patterns, tests segments, and keeps the process more stable.

That doesn't mean SDRs are dead. This debate bores me. Good SDRs with industry understanding are valuable. But SMEs rarely have the luxury of hiring, training, and cleanly managing several of them. If a company like Wittenstein or Schaeffler operates global sales structures, the calculation looks different than for a specialist from Oberderdingen. That's precisely why the comparison must depend on the organizational model.

FAQ: When is AI in Sales worthwhile for manufacturers?

Is AI in sales worthwhile even for small sales teams?

Yes, often especially there. But only if the target market is clear enough. A team of three salespeople can gain enormous relief through AI in sales if account selection, message logic, and handover are clean. If no one can say who the ideal customer is, you automate uncertainty. That then becomes expensive, even with cheap software.

How quickly do you see results?

At MK Kögel, the first reliable responses were visible within the first three weeks, the first qualified appointments in the first month, and a usable segment evaluation after about 60 days. I don't think much of promises after seven days. Technical target markets need learning loops. Anyone who doesn't plan for them confuses activity with sales.

Do you need a perfect CRM beforehand?

No. "Perfect" is a dangerous word anyway. But you need minimum discipline: clear pipeline stages, defined mandatory fields, responsibilities, and clean feedback on what became of an appointment. Without this loop, no AI-SDR system can learn which leads were actually good.

Does this also work in regulated markets like medical technology?

Yes, if you don't pretend you're selling office software. In medical environments, process understanding, hygiene, compliance with standards, documentation, and trust count. A first message doesn't have to clarify all technical details. But it must show that the sender has understood the context. At MK Kögel, precisely this precision was crucial.

Three Learnings from the MK Kögel Story

I take three learnings from MK Kögel that I would also tell other managing directors. Not as a slide. As a warning.

1. Segmentation beats sending volume. If a technical product can serve multiple markets, outbound must be separated by application. Otherwise, responses are generated, but no projects.
2. Sales must guide the AI. Not every answer is equally valuable. A friendly "Interesting, contact me later" is not a sales signal. A brief follow-up question about a special design can be gold.
3. Pipeline is created through rhythm. One-off campaigns are a flash in the pan. MK Kögel benefited because research, outreach, review, and handover ran as a repeatable weekly process.

The second learning sounds trivial, but it's the core. AI can generate text and sort data. It can also recognize patterns. But it doesn't automatically know which inquiry consumes internal effort and which inquiry is strategically valuable. This evaluation must come from within the company. At MK Kögel, it came from sales, application experience, and management. That's why automation didn't turn into a spam machine.

Personal Recommendation: Which solution I would choose

If I were the managing director of a manufacturing SME with products requiring explanation, I wouldn't start with a pure database. I also wouldn't buy an enterprise sales engagement suite first. I would start with the question: Can we describe our target market in such a way that a system generates new, suitable conversations every week and sales then prioritizes better?

For cases like MK Kögel, I clearly recommend an AI-SDR system with operational implementation — i.e., Amplifa or a comparably process-strong solution, if available in the specific market. The reason is simple: The bottleneck is not sending technology. The bottleneck is the translation of technical market understanding into a repeatable pipeline. HubSpot remains important as a CRM and control layer. Apollo or Cognism can provide data components. Salesloft and Outreach are useful if a larger SDR organization is already running. But if you don't have an outbound machine yet, you shouldn't buy a command center.

At MK Kögel, this decision made the difference. The conversations didn't just increase. They became more valuable. After six months, the team could say which target segments delivered real project opportunities and which only produced polite responses. That sounds dry. For a COO, it's music. Quiet music, perhaps. But you hear it in the forecast quality.

Amplifa Sales Audit Check where pipeline is lost in your sales: ICP, data quality, follow-up, CRM handover, and conversion from first contact to opportunity.

Decision Aid: 3 Questions Before Buying a Tool

Before a sales manager books a demo, I would ask three questions. Not to the provider. To themselves.

1. Is our ideal customer profile operationalizable? If you can only say "industrial customers in DACH," you're not ready. If you can name applications, roles, triggers, and exclusion criteria, AI in sales becomes interesting.
2. Who is responsible for feedback and control? An AI-SDR system needs weekly learning. If no one evaluates responses and makes segment decisions, the tool only manages activity.
3. How do we measure success after 90 days? Appointments alone are not enough. Measure positive response rate, qualified conversations, project opportunities, weighted pipeline, research time, and handover quality.

The third question is the most uncomfortable. Many teams want more leads, but no hard definition of quality. Then every provider becomes either a hero or a culprit, depending on who is in the meeting. At MK Kögel, we defined early on what constituted a qualified conversation. That saved discussions. Not all. But the pointless ones.

Why MK Kögel is not an isolated case

I often hear: "Our market is special." Usually, that's true. Trumpf sells differently than Brose, Webasto differently than Phoenix Contact, a sterile goods logistics provider differently than a machine tool builder. But special doesn't mean unstructured. It just means that generic outbound methods fail faster.

MK Kögel is interesting because the company stands precisely between two worlds. On one side, genuine technical consulting, samples, drawings, customer-specific requirements. On the other side, a market that doesn't spontaneously deliver enough qualified inquiries every month. I see this combination constantly. In East Westphalia with component manufacturers. In Franconia with special machine builders. In Baden-Württemberg with medical technology suppliers, where a reception area smells of disinfectant, and yet contribution margins are discussed in the meeting room.

The operational error is almost always the same: Outbound is treated as a campaign. One list, one text, one send. Then people wonder why little happens. At MK Kögel, outbound was built as a process. That's less glamorous. It works better.

The Business Impact: not more leads, but better control

Of course, everyone first looks at the results. 57 qualified conversations in nine months. 21 concrete project opportunities. Approximately 1.8 million euros in additional weighted pipeline. Factor 3.6 compared to the previous outbound pipeline. These numbers are important. They justify the budget.

But the greater effect lay in controllability. Before Amplifa, it was difficult to say whether a weak month was due to the market, activity, the list, or the message. After implementation, the team could distinguish: Segment A responds well but converts poorly. Segment B has less volume but higher project quality. Role C hardly reacts. Role D forwards internally. That sounds like analysis. In reality, it's leadership.

A managing director can make decisions with such data. More capacity in Segment B. Different benefit arguments for Role C. No more time for accounts without a suitable application. This is where ROI is generated. Not in AI itself, but in the decisions that were previously made too late or not at all.

I remember a review in the summer of 2025 where we ended a segment despite a decent response rate. The room was silent for a moment. Printouts of account lists lay on the table, someone had disassembled a ballpoint pen, and a truck drove out of the yard. Then Martin said: "If these don't become projects, it's just busywork." That was the moment I knew: The process had arrived.

What other SMEs can specifically adopt

Not every company has to use Amplifa. I say this deliberately, even though it's on our blog. But every company with technical B2B sales should adopt three things from MK Kögel.

Practical takeaways from the case:

• Build target segments by application, not by industry label. "Automotive" is not an ICP. "Manufacturing sites with sensitive workpiece logistics and high variant diversity" is closer.
• Before sending, define what kind of response is valuable. A demo request, a technical inquiry, a forwarding to the project manager, and a request for later contact do not belong in the same bucket.
• Measure handover quality. Sales needs to know not only that someone responded, but why the account is relevant and what the next logical step is.
• Stop bad segments quickly. Vanity costs pipeline. If a segment shows no project indicators after 200 clean contacts, it needs a new hypothesis or an end.
• Schedule feedback time. 30 minutes per week with sales, marketing, and management can be worth more than 3,000 additional contacts.

The last point is uncomfortable because it demands management time. But that's where the difference is made. Tools without leadership produce metrics. Processes with leadership produce decisions.

Full Success Story The MK Kögel story in detail: initial situation, Amplifa implementation, outbound process, and results from the SME sector.

My conclusion is sharp because the market talks softly enough: Anyone who sells technical products and leaves pipeline to chance is giving away margin. AI in sales is not a substitute for consulting, not a shortcut around market understanding, and not a free pass for bad data. At MK Kögel, it became something more grounded — a pacemaker for conversations that previously arose too rarely. Perhaps that is the most sober definition. And the most useful.