Can SMEs keep up to date with technological developments in AI?
In the new paper in ACM Transactions on Recommender Systems, Hannes Kunstmann, Joseph Ollier, Joel Persson, and Florian von Wangenheim explore how SMEs can successfully leverage developments in Conversational AI in interactions with real customers in the field.
The paper contributes to the developing body of literature by showing that:
- LLM-based systems can deliver strong performance (85.5% recommendation accuracy)
- However, latency (5.7s) and per-interaction costs (~$0.04) remain significant barriers for low-margin contexts
- For stability purposes and consistency in customer interactions, Generative rather than Agentic approaches may be preferable
The project originated as part of Hannes’s Master’s thesis at the Chair of Technology Marketing, ETH Zurich Department of Management, Technology, and Economics.
The full paper is available here: https://dl.acm.org/doi/10.1145/3803546
