AI & digitalisation

Automation in your business: where it has a lasting impact and where it falls short

Moltomedia

Person symbolically buried under sticky notes and memos

*This graphic was partly created with the help of generative artificial intelligence and then edited manually

👉 Let’s start at the beginning

Automation has long ceased to be a topic for the future – it is day-to-day business in production, sales, marketing and, increasingly, administration. And yet studies report two completely different realities in parallel: on the one hand, there are companies that work significantly more efficiently thanks to automation, and on the other, a shockingly high number of projects fail before they can have any effect at all.

The picture is therefore mixed. The question is not whether automation pays off, but where, how and under what conditions.

🖥️ Process automation: the quiet efficiency lever in the background

When people talk about automation, many first think of software that takes over repetitive tasks: invoice processing, preparing quotes, data entry. This is exactly where the biggest and best-documented effects currently lie. An analysis of process automation based on a Forrester study shows an ROI (return on investment) of 383 % over three years with a payback period of just six months. At the same time, according to the same analysis, 78 % of new automation projects in 2025 and 2026 are already relying on AI-augmented approaches instead of classic, rule-based RPA solutions, because AI can also process unstructured tasks such as emails or documents.


Time to escape the old way of working

*This graphic was partly created with the help of generative artificial intelligence and then edited manually


German SMEs present a complex and somewhat contradictory status quo here. Current evaluations of KfW data indicate that around 31 % of German SMEs use AI in their business – even though IT service providers such as Neckar IT advertise an ROI of up to 350 % from process automation alone.

⚙️ Production: automation as the answer to the skills shortage

In manufacturing, automation is no longer an optional extra but a necessity for many businesses. The German Economic Institute (IW) forecasts that Germany will be short of around 728,000 skilled workers in 2027. Besides sales, childcare and nursing, industrial fields such as construction electrics, construction planning and warehousing are particularly affected. The automatica trend index shows accordingly: around half of employees in Germany see robotics and automation as a real opportunity to counter the skills shortage, and 65 % believe that it relieves employees of dangerous or unhealthy tasks.

At the same time, the German robotics and automation industry itself is going through a difficult phase: the VDMA Robotics + Automation association expects sales to fall by 5 % to 14.1 billion euros in 2026, after a 7 % decline the year before. According to the association, however, the long-term drivers – digitalisation, AI, smart production – remain intact. Short-term reluctance to invest and long-term structural necessity are not mutually exclusive here; they simply exist side by side at the moment.

📣 Marketing: automation is becoming a must, no longer a nice-to-have

In marketing, too, the importance of automation is shifting noticeably. The current Bitkom study “Marketing im digitalen Wandel” shows that 76 % of the companies surveyed expect marketing automation to become more important in future, and 84 % see AI as the most important influencing factor for marketing in Germany. At the same time, 54 % name cost pressure and another 54 % the automation of marketing processes itself as their biggest internal challenge – a sign that implementation in practice is often more complex than the trend suggests.

At the operational level, the effects are nevertheless clear: according to practical analyses, small and medium-sized businesses save up to 90 working days a year through marketing automation, such as automated email sequences and lead scoring, and a three-person marketing team can manage tens of thousands of contacts with the right systems without creating additional positions.

⚠️ The other side: why so many automation projects fail

For all the enthusiasm about efficiency gains, one point must not be overlooked: a considerable share of automation projects in German SMEs fail to reach their goal. Current analyses put the failure rate of automation projects in SMEs at around 43 %. The main reasons recur strikingly often:

🎯 Legacy systems without modern interfaces: If old systems don’t provide APIs, every automation becomes an improvised workaround instead of a stable solution.

Imagine this: you want order details to be sent to your ERP automatically. However, the systems you use for this don’t work together that easily, and the only interface is a nightly CSV export for which you use a third system. The automation you want has to bridge this gap; this can be done with either middleware or a custom solution (and in such cases, you can of course always turn to Moltomedia).

🎯 Lack of acceptance in the team: Automation succeeds above all where employees recognise the benefit for their own work and are involved in selecting the processes – not where it is simply imposed on them.

Imagine this: you want to make life a little easier for your employee in accounting and automate the invoicing process, which you consider tedious. The workflow used for this checks incoming payments, generates a PDF, sends it by email and archives it in compliance with the GDPR. Technically, this works within three days. Your employee, however, finds it too risky not to check for themselves what is being sent. For this reason, they check every single invoice again before it goes out. So the automation runs flawlessly, but completely misses its purpose of relieving your accounting employee and taking invoicing off their desk.

🎯 Uncontrolled shadow automation: When individual teams introduce automation tools on their own without IT governance keeping pace, new silos emerge instead of less friction – a pattern very similar to the classic interface problem in companies.

Imagine this: your employees want to work more efficiently and build their own workflows in the systems they use, including Excel macros, and even save them on their private Google accounts because the IT department has no capacity left for it. These workflows run outside IT governance. It hasn’t been clearly communicated which data is processed where; on top of that, nobody has documented the logic, tested edge cases or made sure that IT security requirements are met. If these employees leave your company at some point, access to these workflows in their Excel files is gone in an instant. At that moment, you don’t know whether any data is stored anywhere – and if so, which data and where.

🎯 Wrong prioritisation: Anyone who starts with the most complex process instead of the one with the highest and most easily measurable benefit loses confidence in the entire undertaking early on.

Imagine this: your company uses SAP Business One as its ERP system and Shopware as its online shop, and so far you have maintained your product data in Excel lists that you and your team transfer manually into the other two systems. You decide to solve the “biggest problem” right at the start and launch a fully automated PIM integration that is supposed to synchronise product data, stock levels and prices in real time between SAP, a new PIM system and Shopware. The project affects purchasing, sales and IT at the same time, requires interface development on all three sides and is initially planned to take four months to implement.

A look at the interim status after six weeks reveals … disillusionment … nothing is live yet. Your product data is still being maintained manually in Excel while your team works on the complex solution in parallel. Your sales team notices no difference in their daily work, wonders what the project is actually for and starts to question whether it makes sense.

Yet a simple first step would have been enough: automatically synchronising stock levels between SAP and Shopware – without a PIM, without any pricing logic – could have been implemented within two weeks.

Your sales team would have noticed immediately: no more oversold items, no more manual stock checks. This quick, visible success would have built trust within your team to then tackle the more complex PIM integration – with employees who have already seen from a concrete example that automation really works. (In cases like these, we at Moltomedia are also happy to help with advice and support.)

On top of this comes regulatory pressure: AI-supported automation in a business context is now subject to clear requirements regarding transparency of the decision logic, human oversight of critical processes and traceable documentation. Under the EU AI Act, violations of these obligations for high-risk AI systems can be punished with fines of up to 15 million euros or 3 % of global annual turnover.

💡 What this means for your automation strategy – our Molto tips

  • Start with the process that offers the biggest and most measurable leverage, not the most technically exciting one.

  • Invest in data quality before you invest in tools, because automation amplifies what is already there. If it is based on inconsistent data, your new system will work reliably – but on a flawed foundation.

  • Involve your team from the start. Acceptance often has a far greater impact on success or failure than the software you choose.

  • Think about governance from the outset, especially with AI-supported automation, instead of setting it up afterwards.

  • See automation as a continuous process, not a one-off project. According to current analyses, the greatest productivity gains come from continuous iteration over several cycles, not from the first version of a system.

💙 This is exactly where we come in at Moltomedia

We never look at automation in isolation as a question of tools, but as the interplay of process analysis, system architecture and the people who have to work with it every day. Whether it’s process automation in the back office, interfaces between ERP, CRM and shop systems or AI-supported workflows: we first identify where automation really makes a difference before we talk about specific technology. All under one roof, with just one point of contact and no friction losses between service providers.

Symbolic illustration of humans and AI working together

*This graphic was partly created with the help of generative artificial intelligence and then edited manually

Because the most expensive automation is the one that nobody in the team understands or uses.

Where are you already automating in your company, and where are you still in the middle of the process? We look forward to hearing about your experiences.

🔗 Sources: Bitkom, “Marketing im digitalen Wandel: Zwischen Effizienz, Automatisierung & Wettbewerb” (2026) | abi consulting, analysis of AI process automation incl. Forrester figures on Celonis | Webentwickler.pro, analysis of workflow automation and failure rates in SMEs | GAIM Solutions, analysis of AI automation and KfW data on SMEs | VDMA Robotics + Automation, industry figures 2026 | Packaging Journal, automatica trend index on robotics and the skills shortage | Robotikverband.de, IW forecast on the skills shortage | Neckar IT, guide to process automation and ROI in SMEs | neomarketing.de, analysis of marketing automation in SMEs | EU Artificial Intelligence Act, Article 99: Penalties

Get the clarity you need to move your business forward now.

Start today with your free IT assessment and get a clear roadmap for your digital future in just a few days.

© 2026 Moltomedia. All rights reserved.