Will Virtual Assistants Be The Killer Consumer App for GenAI?

Possible, but not plausible in the near term. Sam Altman believes they are poised to do so (see article from MIT Tech Review.) I've been covering the evolution of virtual assistants as an analyst for more than a decade. I'll share a few thoughts.

Consumer services are always trickier than enterprise ones. Their expectations are higher than those of employees, and their needs are much broader. It isn't just about productivity.

Virtual assistants are an evolution of consumer digital experiences - and they are not a new idea or service. They are nascent (10+ years) and have underwhelmed consumers so far.

Brands have equated virtual assistants or agents with chatbots or voice assistants to date. Chatbots or voice assistants are interfaces that allow us to access information online or control connect devices in our homes (e.g., turning music on/off). To date, they have not been interfaces to intelligent services. GenAI can change that.

Here are a few questions and factors in play that will determine if virtual assistants become the next "killer app" for consumers:

- Smartphones have been morphing into this role for years. They are more than just phones or screens to us already. They orchestrate experiences and services for us. GenAI will enable them to do more. Think evolution here - not revolution.

- Any experience - assistive or agentive - has to be truly convenient. Ever heard someone say, "ah, forget it. I'll just do it myself." Sometimes it really is easier just to do things by yourself than to explain to someone else (or a machine) how to do it for us. Conversations aren't the easiest way to do everything - even with "eyes" or video/images to assist. Invisible experiences (i.e., whereby assistants anticipate consumer needs) will be even more convenient.

- Consumers need to trust assistants or agents. This will happen over time as these new services protect consumers' privacy and deliver on the promise of convenience (mostly). Think evolution rather than revolution in terms of pace. Consumers will trust assistants or agents over time as they deliver the value they've promised.

- Assistants need to know us. How will they get to know me? Apple and Google know me really well because I have been using their digital services for nearly two decades. Will this history matter? Or will I find a way (and trust the agent) to unload "my life" into an application (LLM-based) such that it can understand me and assist me nearly instantly? Will there be vertical versions?

- Consumer hardware and networks must support the services. What will be the architecture and infrastructure (i.e., how much on our devices and how much in the cloud?) Consumers don't like latency. Imagine how mobile technology has reset our expectations. Consumers don't want to swipe and tap their phones for 15 seconds to get a ride, buy a ticket, or order take-out. They want it to be even faster.