Heavy on Marketing, Often Light on Substance: How AI Creates Real Value in Your Company

Der Markt für Künstliche Intelligenz (KI) boomt. Immer mehr Anbieter preisen ihre Lösungen an. Doch bei genauer Betrachtung entpuppt sich so manche vermeintlich intelligente Neuigkeit als Mogelpackung, schließlich lässt sich allein mit dem Label „Künstliche Intelligenz“ ordentlich Geld verdienen. Mittelständische Unternehmen stehen dadurch oft vor einem Dilemma: Einerseits möchten sie KI-Lösungen einführen, um beispielsweise von Effizienzsteigerungen zu profitieren, andererseits ist die Auswahl von Tools aufwändig und bleibt oft ohne Ergebnis. In ihrem Beitrag erklären die enomyc-Autoren Carla Dausend und Leonard Rampf, welche Fehler Unternehmen vermeiden sollten und was beim Einstieg ins Thema zu beachten ist.

Four typical pitfalls companies should avoid – and what really matters instead

These four pitfalls are particularly common:

1. Focusing on technology instead of problems Many companies begin their AI journey by searching for the “best AI” instead of asking themselves: What problem are we actually trying to solve? The result: Solutions are implemented that do not meet actual needs. A reverse approach is more promising: start with the specific business problem and only clarify in a second step whether AI can make a meaningful contribution. Recommendation: Start with a structured use case workshop in which operational pain points and data potential are systematically identified.

2. Tool fetishism instead of a target vision Whether chatbots, recommendation engines, or predictive maintenance – in many companies, the discussion is dominated by the choice of tools. But tools are only a means to an end. If you don't formulate a clear target state, you can easily get lost in endless comparison tables and test phases. Recommendation: First develop a target vision for processes and results. What exactly needs to be improved? How do we measure success? Only then should the question of technology be addressed.

3. Lack of customization An off-the-shelf solution rarely fits perfectly. Many companies purchase AI products that are not compatible with their own data structure or business model. The result: implementation effort, low acceptance within the team, and disappointed expectations. Recommendation: Ensure that technology partners work closely with your specialist departments. Pilot projects help to check in advance whether the solution fits the specific requirements.

Our approach: expertise, clarity, and genuine solution development

Best practices from our projects

Three typical examples:

1. Logistics: A retail company with its own fleet of vehicles plans its routes manually – current factors such as sick leave, traffic, or delivery priorities are not taken into account. Our solution: An AI-supported route module with proven logic, supplemented by company-specific requirements. Added value: Up to 15 percent fewer kilometers driven per delivery and significantly greater schedule reliability in the event of short-term cancellations.

2. Sales: Many CRM systems have untapped potential. We help identify relevant cross-selling opportunities based on existing sales data – with a modular analytics approach that can be tailored to specific product ranges and purchasing behavior. Added value: An average of 5 to 10 percent more revenue per active customer relationship and a significantly higher success rate for sales campaigns.

3. Purchasing: Many risks in the supply chain go unnoticed – until it's too late. Our system automatically analyzes dependencies on strategic suppliers and links this information to external data sources such as news or creditworthiness information. Noteworthy constellations are identified at an early stage. The logic behind this is standardized, while the evaluation remains individually controllable. Added value: Risks are systematically screened and identified at an early stage, resulting in up to 25% fewer delivery failures and significantly fewer (expensive) ad hoc procurements.

Fewer buzzwords, more impact