The Virtual Assistant market is in a period of change: “traditional” intent-based platforms are being retired in favour of new solutions underpinned by large language models (LMMs) and generative AI. Development had stalled somewhat in recent years; this was not due to a lack of innovative technology but instead because user engagement remained locked into
simple but non-monetizable “command and control” based functions. Generative AI is, therefore, a welcome disruption in the assistant landscape, promising a reset of capabilities.
Advancements in large language models are now a vital constituent in delivering more conversational and personalised virtual assistant platforms. AI has enabled the addition of intonation and emotion in verbal responses, conveying mood and other non-verbal information to users. Indeed, all leading assistant platform vendors are now following broadly the same innovation pathway, with large language models providing the catalyst to design more intelligent voice interfaces.
Meanwhile, a parallel activity in developing “small” language models (SLMs) – those optimised for on-device tasks – looks likely to expand the opportunity to deliver more cost-effective solutions overall.