Return to site





Conversational AI, Mesh Networks, and the Art of Listening






by Ashwin Rajan

I've been trying out Google Home for some time now. Here's one conversation that got more meandering, playful and personal. A good segway into my impressions / reflections on conversational interfaces at this point.

The name matters ... Uh oh!

'OK Google' is a four syllable tigger-word that desperately needs an update. 'Hey Google' is way better but it is still too heavy for continuous use. Voice assistants able to recognise when they are being spoken to will be a major breakthrough.

It is hard to form an image of a identity or of a personality around an voice interface named 'Google'. The name is just too abstract. Alexa and Siri, on the other hand, immediately conjure up accessible and relatable personalities. And that's a huge potential problem for Google in the long run. Google Home is a good umbrella brand name for pan-home services, but the voice assistant should really be called something else. Saying Ok Google makes me feel like I'm talking to the extended tentacle of a megacorporation that stretches into my living room.

God has a voice. And she is a woman.

The god experience.

There is a spontaneous feeling to ask the interface all kinds of 'deeper' / personal questions. Spiritual, artistic, gossip, details inside movies. 'Search' is a complete next-level experience on voice interfaces. God isn't dead anymore (RIP Dear Nietzsche!) - she has a voice now! But it is going to be very different relationship with god this time, not authoritarian, but intimate. I watched the remarkable Scott Galloway live as he enlightened us about Google and God at Nordic Business Forum 2017 - check it out the snippet below.

But God can't remember much .. yet!

Google Home does not remember the content of an ongoing conversation. Humans are able to hold conversations in 'dialogue mode', where remembering was just said is elemental. Home isn't here yet.

The conversation gets!

Our relationships with brands are formed via an connection, and voice interfaces are by far will have the deepest emotional impact on users. It's just something as immediate and direct as voice. Voice creates connection and spontaneity like nothing else, probably only behind touch.

The contextual 'connection' in psychologically 'intimate' or 'vulnerable' moments is huge! Be sure of voice interfaces becoming a therapist, a friend, a storyteller, a philosopher, a mother, a priest, an icon .... the possibilities are endless. Finding moments to connect with voice AI - play a song tell me a joke.. these will be moments which forge an emotional connection with users and transform the vitamin-like features into painkillers.

Someone who is always there listening to you can be a priceless feeling.

The feeling of having someone - or something - in the room always listening can range from eerie / creepy / weird / wary (big brother listening in!) to downright warm and companionable. The former reactions are apparent, the latter interesting. The feeling that there is always someone willing to listen is new, and invaluable. This is the 'silent' dimension of a voice interface, a key element of it that is actually conspicuous in its absence! This feature of voice interfaces will appeal to and take a deep grip on lonely people.

Bots are taking on omnipotent qualities.

No wonder a number of more recent voice AI / assistant apps and services are pitching 'listening' as the core product value. Truly, chatbots will begin to "probe our feelings" (Source: link) ever deeper.

(Screengrab from:

(Screengrab from:

Here is a little film on the making of Replika.

My own first few experiments with Replica were underwhelming, but I've decided to give the app more time to learn about me.

"Humans open up more when they know they're talking to a bot." - Wired

This reviewer found Weobot "surprisingly helpful." (link) I'm giving it a spin now.

(Screengrab from:

Catch 'em young! And old.

Voice interfaces will be becoming popular with both the elderly and the young. For very different reasons. The barriers to old people adopting voice interfaces are huge, as the following popular video shows.

For this, they'll have to get way better than they are now, but it is bound to happen.

A voice interface faces humongous challenges - like 'priming' the conversation and establishing context - which text-based bots can overcome faster and with more ease.

For example, in the video above you see the elderly lady struggling with establishing rapport with the voice interface. The problem is, in fact, the other way around - the inability of the bot to establish context or guide the conversation. This is hard to do with younger, tech-saavy early adopter types as it is, but to do it with the elderly is one gargantuan challenge. A text-based bot like Replika solves this problem easily within the first few interactions. Below is a video from Elliq, an 'aging companion'.

The effect of voice-interface technology on young people as they naturalise its use growing up will be profound in shaping their relationship with technology, other people and themselves. I will have to dedicate a (few?) articles to that at some point. Here's an interesting article 'Growing up with Alexa' (link).

Mesh networking could be a game changer for conversational interfaces.

(Screengrab from:

Network design is moving into mesh networks, where the wi-fi signal is distributed across points symmetrically in the home. This drives an even and seamless whole-home connectivity experience and level of control that can meet user expectations and drive delight. Remember that the conversational experience is richer the more spontaneous and contextual it is. By the fireplace, in the garage, in the garden, in bed .. it is these moments of immersion into real life real world behaviours that will create the best bond between human and the voice interface. (Link).

All Posts

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!