As organisations rush to adopt AI agents, the urge to use them everywhere can result in over-complicated solutions. We spoke with Taras Firman to find out when agentic AI really adds value and when other methods work better.
Background & experience:
- More than a decade of experience spanning computer vision engineering, mathematical modelling, data science, and consulting.
- Focuses on Explainable AI, addressing the critical challenge of building trust in AI models for both academic and business environments.
- Mathematician by training with a Ph.D. in Mathematical Sciences and extensive cross-industry experience across retail, logistics, banking, healthcare, and bioinformatics.
AI agents have become one of the hottest topics in tech. What makes them such a compelling paradigm for building intelligent systems?
Taras Firman: AI agents can break down complex tasks into smaller, sequential steps and use dynamic tools, making them useful in many real-world cases. However, they are not a universal game-changer and should not be applied everywhere by default.
Often, simpler and more stable methods like deterministic workflows, classic ML models, or even single LLM calls are more reliable, cheaper, and easier to maintain. The key is to use agentic AI where autonomy, multi-step reasoning, and adaptability really matter, rather than adding needless complexity.
How do teams figure out in advance whether an agentic approach is the right fit for a given use case?
TF: It’s simpler than many believe. AI agents work well for tasks that are naturally complex and need adaptive decision-making across multiple steps or sources of information. When it is needed to make sequential decisions or process different sources of data. Another sign can be situations where a system must reason, plan, and dynamically decide what to do next, rather than just executing a predefined set of instructions.
For example, an agent can coordinate various tools and data sources to help a business analyst prepare a report. It might collect data from multiple databases, gather external market info, clean and summarise it, then create a detailed report. In IT support, an agent can troubleshoot issues by checking logs, querying systems, suggesting fixes, and asking questions if needed. In these cases, the agent’s flexibility and independence clearly outperform simpler workflows or single-model methods.
So where do agents fall short? When should teams steer clear of the agentic approach?
TF: Agentic AI is often overkill for tasks that are predictable, well-defined, or should be data-specific. Low-level operations, like classifying emails as spam, validating input in a web form, and converting file formats, don’t get any benefits from an agent’s planning or reasoning capabilities. Even tasks like checking compliance against fixed rules are better handled with deterministic logic or a simple ML model. Adding an agent to these tasks only increases complexity, latency, and the risk of errors, without delivering meaningful benefits.
Given all of that, what does a well-designed system look like in practice?
TF: In practice, the best results often come from combining agentic AI with simpler, deterministic or ML-based components. Agents are great at high-level reasoning, planning, and deciding next steps, but they can be slower and use more resources. Pairing them with low-level, specialised models or rule-based workflows for predictable tasks like data validation or classification offers the best of both worlds: speed and reliability. This hybrid approach boosts performance and efficiency while reducing errors and operational overhead, making agentic systems truly practical for real-world applications.
Related insights
The breadth of knowledge and understanding that ELEKS has within its walls allows us to leverage that expertise to make superior deliverables for our customers. When you work with ELEKS, you are working with the top 1% of the aptitude and engineering excellence of the whole country.
Right from the start, we really liked ELEKS’ commitment and engagement. They came to us with their best people to try to understand our context, our business idea, and developed the first prototype with us. They were very professional and very customer oriented. I think, without ELEKS it probably would not have been possible to have such a successful product in such a short period of time.
ELEKS has been involved in the development of a number of our consumer-facing websites and mobile applications that allow our customers to easily track their shipments, get the information they need as well as stay in touch with us. We’ve appreciated the level of ELEKS’ expertise, responsiveness and attention to details.
