AI agents are becoming one of the most talked-about innovations in enterprise technology. Companies are exploring ways to integrate AI-driven automation into their operations, but are AI agents truly transformative, or are they just another phase of hype?
In a recent episode of the CLARITY.SHOW podcast, host Karyna Mihalevich, a Chief Product Officer at CLARITY and SAP Intelligent Enterprise Ambassador, sat down with Lasse Rindom, AI Lead at Basico, to discuss the real impact, challenges, and future of AI agents in business.
Understanding AI agents: more than just a buzzword
AI agents are often positioned as the next frontier in enterprise automation. But what exactly makes an AI agent different from traditional automation? As Lasse Rindom explains: “If we boil it down, an agent is something proactive, capable of acting on its own.”
He emphasizes that while AI agents can handle complex, unstructured tasks, not every business process requires such sophistication. “Most things we do in companies are deterministic, structured processes. Turning your entire enterprise into an agile agent is over the top,” he says. This is a crucial insight for businesses: AI agents should be deployed where they can add real value, not just because they are the latest trend.
The challenges of AI implementation
One of the biggest challenges companies face when pursuing AI is the hype surrounding it. Lasse warns against blindly following trends without a clear understanding of the technology’s limitations. “You need to map out the complexity to test it. But if you’re mapping it all out, why not just do something deterministic?” he questions. Additionally, testing takes much longer because AI agents don’t just execute commands – they infer, learn, and sometimes even ‘hallucinate’ responses, Lasse points out.
Karyna adds to this by noting the extended testing cycles required for AI agents. “The testing cycle is longer than in traditional development because it’s partially a black box for us,” she explains. This unpredictability is a significant hurdle, especially when AI agents are given access to critical systems.
The role of AI in human-AI collaboration
As AI agents become more common in enterprises, the question of how humans and AI should collaborate becomes increasingly important. Lasse advocates for a “human in the loop” approach, where AI supports human decision-making rather than replacing it entirely.
“We need to start having discussions about what are the problems and the solutions we want to have. Don’t automate your interactions with customers if your value proposition is to talk to them,” he advises.
Rindom also uses a vivid metaphor to describe the kind of AI agents he envisions for enterprises. He doesn’t want AI agents to be like James Bond, acting autonomously and unpredictably. Instead, he prefers AI agents to be more like performers on a stage: “I want an agent that will go on stage, perform the thing I asked it to perform, and then go down again.”
This metaphor underscores the importance of assigning specific roles to AI agents, rather than giving them broad, uncontrolled autonomy. “I want to have an agent that will perform a role, not one that is proactive with a license to kill,” he adds.
Karyna agrees, emphasizing the need for a human-centric approach. “We always should put humans first and our needs first. This is very important when implementing AI tools, whether in enterprises or in our personal lives,” she says.
Selecting the right AI tools
With a growing number of AI tools available – ClickUp AI for project planning, Fireflies.ai for meeting transcription, GitHub Copilot for code assistance – how should companies decide which to adopt?
Lasse advises businesses to focus on their existing ecosystem rather than chasing new tools: “Don’t start with AI. Start with your existing systems. If you’re running on-prem legacy systems, you won’t benefit from AI in the long run. Modernization should be your first step.”
He also cautions against relying solely on vendor promises: “Software providers are now branding everything as AI-powered. We need to move beyond the buzzwords and ask: Does this tool solve a real business problem?”
The future of AI agents in business
Looking ahead, Lasse predicts that AI agents will continue to evolve, but businesses should focus on solving specific problems rather than chasing the latest trends. “We need to make sure that AI brings value where we’re using it concretely and that we can maintain the things we’re building,” he says.
The AI expert also reflects on the broader implications of AI agents, using another metaphor to describe the current state of AI development. He compares the introduction of AI agents to “James Cook arriving in Australia and seeing kangaroos for the first time.” Just as Cook had to figure out how to interact with these unfamiliar animals, businesses today are still trying to understand how to integrate AI agents into their operations. “We are not coming to a completely new continent, but we are still figuring out how to treat this new ‘animal’ called AI,” he says.
Karyna concludes by highlighting the importance of experimentation and creativity in AI adoption. “Having new tools opens up a level of creativity and imagination for employees. We can explore new areas and find additional value by using these tools,” she says.
Conclusion: practical AI, not just AI for AI’s sake
AI agents have the potential to reshape enterprise operations, but only if implemented with a clear strategy. As the podcast discussion revealed, the focus should not be on the AI itself, but on how it enhances business processes, supports human employees, and delivers tangible value.
Companies that take a measured approach – asking ‘why’ before ‘how’ – will be the ones that truly benefit from AI, rather than getting lost in the latest wave of technological hype.