AI Models (SLMs and LLMs)
What are AI Models (SLMs and LLMs)?
The Upstream Works AI Experience Platform orchestrates specialized, economical Small Language Models (SLMs) and private Large Language Models (LLMs) purpose-built for customer experience and contact center operations. It orchestrates the optimal mix of AI models and matches the right model to each task to maximize performance while maintaining governance.
Contact-center AI rarely relies on a single model. Small language models (SLMs) are compact models tuned for narrow, repeated tasks such as intent detection, data extraction or summarization, and they are fast and inexpensive to run. Large language models (LLMs) are broader and better at open-ended reasoning and natural responses, but they cost more per request and need stronger controls.
Orchestration means routing each task to the model that fits it, balancing accuracy, latency, cost and data privacy. Running models privately helps keep customer data inside approved boundaries, and grounding responses in approved knowledge reduces the risk of hallucination. When evaluating an approach, buyers typically look at task-level accuracy, response time, cost per interaction and how model behavior is governed and audited.
Related: Small language model (SLM), Large language model (LLM), Grounding, UWXP.