Home
DE
Expert Articles
Discover well-researched expert articles on the key developments in artificial intelligence — from agentic AI and computer vision to the EU AI Act and industry-specific applications. The articles combine technological context with concrete practical examples and a clear focus on measurable value.
Agentic AI: When AI doesn't just answer, but gets work done
Many AI applications stop at well-crafted answers. Agentic AI goes one step further: it receives a goal — for example, “Create a complete quote for customer XY” — and works toward it autonomously: it plans substeps, calls APIs, reads databases, executes code, and checks intermediate results until a reliable outcome is produced. The principle: goal → plan → action → check → adjust → result. Instead of just producing text, this AI acts like a digital employee.
Goal Orientation
The system works toward a clearly defined result — e.g. a finished report, a completed booking, or a validated data object
Planning & Decomposition
A goal like “Analyze quarterly data and create an executive summary” is automatically broken down into 5–10 substeps and processed sequentially
Tool Use
Access to REST APIs, SQL databases, web browsers, Python interpreters, email systems, CRM, and ERP — depending on the task
Feedback Loop
After each action, the system checks the result against the goal: Does it match? If not → correction and another attempt
Context & Memory
Intermediate results, decisions made, and failed attempts are stored in persistent working memory (e.g. a vector store)
Control & Guardrails
Policies define which tools may be used, which actions require human approval, and when the process should be stopped
Classic Chat AI
  • Reactive: question → answer
  • Usually a single output
  • Hardly any autonomous actions
  • Limited to text generation
Agentic AI
  • Active: goal → process execution
  • Multiple iterative steps
  • Uses tools and checks results
  • Delivers a verifiable outcome
Typical enterprise use cases
Business Automation
Automatically generate quotes from CRM data (Salesforce, HubSpot), analyze RFPs, and create proposal drafts, fully run onboarding workflows for new customers
Data & Analytics Agents
Pull raw data from data warehouses, clean it, run models, and output a management summary with recommendations in natural language — without manual intermediate steps
IT Ops / DevOps
Analyze logs from Splunk or Datadog, triage incidents by severity, execute runbooks automatically, and create escalation tickets in Jira when needed
Research Agents
Search patent databases, scientific publications, and competitor websites, extract relevant content, and turn it into a structured research report — in minutes instead of days

Key takeaway: Agentic AI = a digital employee with a goal, a toolbox, and self-control. It doesn't just produce text — it completes tasks from start to finish, in a verifiable and traceable way.
What really matters in practice: Guardrails
To make agentic AI productive, a clear security and governance design is required. These safeguards are not optional; they are essential for safe operation:
Human-in-the-loop
Actions with external impact (sending email, deleting files, triggering payments) require explicit approval
Least Privilege
Each agent receives only the minimum access rights necessary — no write access where read access is sufficient
Logging & Audit Trails
Every action, every tool call, and every decision is logged with timestamp and rationale — GDPR-compliant
Rate & Budget Limits
Maximum API calls per hour, token budgets, and cost limits prevent uncontrolled spending or endless loops
Termination Criteria
After N failed attempts or when confidence falls below a threshold, the agent stops and escalates to a human
Advantages
  • End-to-end execution: A process like “analyze RFP → create proposal → send” runs fully automatically — time savings: 60–80% compared to manual processing.
  • Robust in multi-step workflows: Self-correction in case of errors reduces manual rework by up to 70%.
  • Scalable: Hundreds of parallel agent instances handle different tasks simultaneously without quality loss.
  • Consistent data quality through automated validation and cleansing at every process step.
  • Employees focus on exceptions and strategic decisions — the AI handles routine work.
Risks to keep in mind
  • Hallucinations in tool calls: The model calls an API with incorrect parameters — result: faulty data or unintended actions. Countermeasure: output validation before every action.
  • Permission creep: Overly broad access rights enable unintended data access. Countermeasure: least-privilege principle and regular rights audits.
  • Compliance risks: Automated data access must be GDPR-compliant — audit trails and a data protection impact assessment are mandatory.
  • Overreach: Without clear stop criteria, the agent performs actions that were not authorized — e.g. sending emails to external recipients.
How Jaroona supports this
Jaroona develops agentic AI systems that integrate directly into existing corporate infrastructure — from SAP and Salesforce to internal databases and cloud platforms. We support the entire journey: from process analysis and feasibility study to proof of concept and production-ready solution with full monitoring. The focus is on four core areas:
Operationalization
From idea to production-ready solution: we define process boundaries, build robust error handling, and ensure the agent remains stable even with unexpected inputs — with SLA monitoring and automatic retraining.
Tool Integration
We develop customized tool adapters for your systems: REST APIs, SQL databases, SAP modules, Microsoft 365, Salesforce, internal knowledge bases, and proprietary interfaces.
Governance & Guardrails
We implement role-based access models, human-in-the-loop workflows for critical actions, complete audit trails, and GDPR-compliant data storage — documented and auditable.
Measurability
We define KPIs before go-live: turnaround time, error rate, cost savings, automation rate. Dashboards in Power BI or Grafana make ROI transparent and continuously traceable.

With Jaroona, you transform concrete business processes — not just pilot projects. Our agentic systems run in production, are secure, and deliver measurable ROI.
AI-powered analyses for drones are transforming inspections
Jaroona develops vendor-agnostic AI software that integrates directly into existing drone systems—without any hardware replacement. The solution combines high-resolution 4K/8K RGB cameras, multispectral and thermal sensors, as well as LiDAR technology to generate precise 3D terrain models. Deep learning models based on YOLO and CNNs process image data in real time via edge computing directly on the drone—inspection times drop by up to 70%, and detection accuracy reaches over 95%.
The AI delivers georeferenced, reproducible findings with a resolution of up to 1 mm/pixel. Results are available both as real-time feedback during flight and as a structured cloud report—including automatic prioritization by urgency. Cost savings compared with manual inspections: up to 40%.
Energy infrastructure in focus
Our AI systems are specifically optimized for inspecting high-voltage power lines, wind turbines, photovoltaic parks, and dams. For dams, data is captured at 1 mm/pixel resolution and evaluated using specialized segmentation models (U-Net, Mask R-CNN) that distinguish between crack types, material changes, and moisture ingress.
Cracking & structural changes
YOLO-v8-based object detection identifies hairline cracks as narrow as 0.1 mm and corrosion on concrete and steel structures. Detection rate: over 98%. Findings are automatically classified (severity levels 1–4) and tagged with GPS coordinates.
Erosion & moisture zones
Multispectral (NIR, red edge) and thermal cameras capture temperature anomalies and moisture gradients within the dam body. AI models detect seepage and erosion zones invisible to the human eye—up to 3 weeks earlier than conventional methods.
Anomalies in dam bodies & structural joints
Deep learning segmentation models analyze surface irregularities, settlement cracks, and the condition of expansion joints. Deformation analyses are performed by comparing georeferenced 3D point clouds from successive inspection flights.
Other use cases for drone AI
Wildlife monitoring & species conservation
YOLO-based detection models identify and count wildlife using 4K/8K RGB and thermal cameras with an accuracy of over 95%—even in dense vegetation or at night. Compared with manual counting methods, this saves 60–80% of time and enables detailed population dynamics analyses for informed conservation strategies.
Plant & vegetation analysis
Multispectral (NIR, red edge) and hyperspectral cameras capture NDVI values and other vegetation indices. CNN models detect plant species, nutrient deficiencies, water stress, and pest infestations with over 98% classification accuracy—and reduce fertilizer and pesticide usage by 15–30% through precise, need-based application.
Soil conditions & erosion management
Photogrammetry and GPS-based aerial imagery generate high-resolution DEM models (Digital Elevation Models). AI algorithms model erosion zones, slope instability, and soil erosion with over 92% accuracy—and reduce project costs for soil protection measures by 15–25% through efficient data capture.
Solar park & wind turbine inspection
Edge AI on inverters and drones detects hotspots, faulty modules, and cable issues in real time. For wind turbines, the system analyzes vibration and temperature sensor data to identify bearing and gearbox damage early—unplanned outages drop by up to 50%, and plant availability increases measurably.
Technology stack at a glance
70 %
faster inspections compared with manual methods
>98 %
detection rate for cracks, corrosion, and structural damage
40 %
cost savings through automated analysis
3 wk.
earlier detection of moisture damage compared with conventional methods

All finding data flows automatically into a structured inspection report: prioritized defect list, severity classification, GPS coordinates, and photo documentation—directly exportable to common maintenance management systems (CMMS) such as SAP PM or IBM Maximo.
Mastering the requirements of the EU AI Act
According to the EU AI Act, which aims to build trust and acceptance for AI in the European Union, a high-risk AI system is one that carries a significant potential to negatively affect the health, safety, or fundamental rights of individuals. This applies in particular to systems used in safety-critical areas or that make critical decisions with far-reaching impacts on people, as set out in detail in Annex III of the Act.
Here is the official and practice-oriented overview of which AI systems qualify as “high-risk AI” (according to the final EU AI Act, adopted in March 2024, in force since 2025):
Categories of high-risk AI
Biometric identification and categorization
AI systems for remote biometric identification of persons in real time or ex post (e.g., facial recognition in public spaces or for workplace surveillance) and for categorizing persons based on biometric characteristics (such as gender, ethnic origin, sexual orientation, or emotional states).
Management and operation of critical infrastructure
Systems used for the operation or management of critical infrastructure such as electricity, gas, and water supply, traffic control systems, digital infrastructure, or medical emergency services, where failure or malfunction could significantly endanger the life, health, or safety of people.
Education and vocational training
AI systems intended to determine or evaluate access to educational institutions (e.g., admission procedures), assess learning performance (e.g., automated exam systems), or assess the suitability of individuals for specific educational paths or vocational qualifications, which can affect career opportunities.
Employment, workforce management, and access to self-employment
AI systems used for recruiting and selecting personnel (e.g., résumé filtering, psychometric tests), evaluating employee performance, promotion or dismissal, assigning tasks, or monitoring work behavior, as these can directly affect working conditions and livelihoods.
Requirements for high-risk AI
1
Robust risk management process:
Implementation of a systematic approach to identifying, analyzing, assessing, and mitigating risks throughout the entire lifecycle of the AI system, from development to disposal.
2
High data quality and technical documentation:
Ensuring the quality of the training, validation, and test datasets in terms of relevance, representativeness, and error-free content. Creation and maintenance of comprehensive technical documentation that enables transparent review of the system.
3
Transparency obligations and human oversight:
Ensuring that the operation of high-risk AI systems is understandable to users (“explainability”). In addition, effective human oversight must be possible at all times so that unwanted outcomes can be corrected or the system can be shut down.
4
Conformity assessment and CE marking:
Before a high-risk AI system is placed on the EU market, it must undergo a strict conformity assessment procedure. After successful completion, it receives the CE marking, confirming its compliance with all EU requirements.
5
Registration in the EU database:
Mandatory registration of all standalone high-risk AI systems in a public EU database to ensure transparency and facilitate monitoring by market surveillance authorities.
Examples for classification

Jaroona offers comprehensive expertise in the correct classification and assurance of compliance for your AI systems in accordance with the complex requirements of the EU AI Act and supports you in implementing the necessary risk management processes.
Stay informed
Would you like to stay up to date on the latest developments, publications, and events from Jaroona? Subscribe to our newsletter or follow us on our social media channels.
We regularly share valuable resources, expert articles, and news from the world of AI and cybersecurity. Don’t miss any important updates!

t6517e3d4.emailsys2a.net

Newsletter-Anmeldung

Mit unserem Newsletter verpassen Sie keine Neuigkeiten mehr. Jetzt anmelden und auf dem Laufenden bleiben!

Your Partner for AI & Machine Learning
Tailored AI solutions for businesses — from strategic consulting to successful implementation.
© 2026 Jaroona. All rights reserved.