Semi-sentient, voice-first, dynamic websites are (not) coming anytime soon

One of my amazing colleagues wrote an article on a new form of Careers Website that is ready to launch right now (at least according to this article).

Now let me be super clear: We produce some of the best Careers sites in the market using the tools/tech that is currently available in the wild. We produced one of the first Careers chat bots, but we won’t push chatbots if that isn’t going to solve your business problems. We use enterprise search insight engines to analyse how candidates search to find content/jobs and use that insight to improve the journey for future candidates.

I’m on the Innovators section of the technology/innovation adoption lifecycle bell curve, but for most clients we push them towards the Late Majority side of things. That’s not to say that they are late to the party for the recruitment solutions, far from it, but the recruitment industry is behind the consumer industry… so late majority as a eco system rather than industry. I try to keep our most ambitious clients between the Early Adopters and Early Majority side though, and I’m glad to have a few of those clients that like to push the boundaries.

Right, now that’s cleared up let’s look at the key idea.

The semi-sentient dynamic website concept

The idea is that you visit the website, enable your video and microphone, and start asking questions.

Using the words that you say the website will dynamically present you with the content that you require to answer your question. In addition to that, the website will also use the expression of your face, along with the tone of your voice to understand your sentiment and adapt the way in which it delivers that contnet back to you (and likely the type of content that it delivers back to you).

Sweet concept, and potentially the future of how websites will exist (with added emphasis on potentially).

The section where I get pessimistic/realistic

Let me go into some of the reasons around why I’m this stuff is not ready for prime time and say that this stuff is a good 2-3 years away from being more robust and more of an opportunity to execute.

Content Creation

When you create a website you define a sitemap and information architecture. These things are often based on the type of content that they user/candidate needs to know to be able to

  • Video requires a solid internet connection that won’t fail
  • Audio requires a solid internet connection that won’t fail
  • Accents or candidates that don’t understand English quite a well (a friend of mine from Poland can’t listen to my podcast because he finds it hard to understand my accent, but when he speaks with me in person it’s fine…. Weird right. For that reason I used to transcribe all the podcasts for people in similar situations as well as those who have hearing difficulties).
  • The responses need to be preprogramed. ALL of them. Not only would you need an answer to questions about the company culture, but you would need it to be answered in several different ways with different inferences to hit the right response that the candidate would be expecting.

I’d be interested to know what you think about the statement that “Candidates would love it” as well. The article goes to explain that we’ve been speaking for hundreds of thousands of years, but written communication has only been around for 5000. In today’s age it seems to me that people MUCH prefer to receive video/image content (Instagram, youtube, tiktok, snapchat), but that when you’re talking with people it is done over short written messages (sms, imessage, snapchat, insta, twitter, what’s app).

It’s also really uncomfortable and not socially acceptable (yet) that people will speak out loud to their devices and to get answers back. For one, I feel like an idiot, but also I don’t necessarily want people to know what I’m doing on a careers site if I’ve already got a job somewhere.

A couple of years ago I was doing a talk on the future of the web and covered off VR and future interactions. I was trying to demonstrate how you would navigate a website using a VR headset and found it really frustrating that you had to use a virtual keyboard to type in search options, so I put a voice search on the site : https://responsivedesign.is/speech/ (use the Chrome browser if you’re testing this, it’s still in beta). I was super self conscious about using it at work and never used it in public, but it also had major issues with understanding my accent… damn Australian’s.

Then there’s the whole collecting of data points and storing them (the video, the audio etc)…. A GDPR nightmare ☹

Anyway, I think the concept is amazing and the potential for all these things to come together is certainly there…. but in practice we’re still a way to go.

Now, this article was originally written 12 months ago… I just forgot to press publish. Now, with the arrival of GPT3 we might have the nuances of AI to produce the right content given enough training…. what do you think?

Leave a comment

Your email address will not be published. Required fields are marked *