Say Hello 👋

Last week Facebook announced the closed beta of its new virtual assistant, M.  Working on top of its messenger app, M is pitched squarely as a step beyond products such as Siri, Cortona or Google Now. The big difference is that unlike its competitors, M isn’t supported by artificial intelligence alone.  Facebook has opted for a hybrid approach – using AI to answer straightforward questions and employing real people – M trainers – to help answer more complex, nuanced requests.

For anyone that has used Siri for one intention only to be given the answer to something completely different then it’s fair to say that a true AI-based solution is some way off, and Facebook with M has acknowledged that.

It’s an admission that whilst AI is the future, it’s not the ‘now’ for users.

To that point, we have seen a number of SMS-based virtual assistants crop up in the US, most noticeably Magic , the personal assistant that promises to get you anything you want (as long as it’s not illegal) and others such as Go Butler

Whilst Apple, Google & Microsoft continue to invest and iterate at pace in this space it’s clear that they believe this is the future but the blend of AI with human intervention of M will almost certainly see a much smoother user journey, more satisfied users with a much higher retention rate than any of the others.

Therefore, as their AI engine improves M will seamlessly be able to phase out the human intervention (and attributable cost) without harming the end user experience.

The other notable feature of M is its ability to complete tasks for the user like buying items, delivering gifts or booking travel tickets. – adding a deeper utility value to the whole experience.

There is no timescale at present for M to be launched to the masses but for once, I think one of the big guns may have nailed an approach to virtual assistants that puts the user, context and usefulness right at the heart of the experience.