AI – Separating the Wheat from the Chaff

AI has been at the center stage over the last couple of years. Conversations around the adoption of AI across several use cases, the possibilities of AI replacing/augmenting human workforce, the readiness journey of enterprises to integrate AI into their product mix, the democratization of AI beyond just data science teams, social contracts with conversational AI, AI vs. emotional intelligence, and several such topics have consumed tech and marketing discussions.

In technical terms, deep reinforcement learning, NLP, and generative modeling made rapid strides in 2018. There were lots of AI conferences, lots of developer activity on GitHub around this mega trend, and a lot more AI related courses in schools.

As with any tech on the curve, some of AI is hype, but some of the adoptions have been real. Some AI as has crossed the chasm, and some still waiting for early adopters. Clarifying what hype is - it is 'talk' about a future state when the current state exhibits adoption challenges. The challenges are real and may exist for various reasons - gaining talent, transforming legacy systems, ingraining the DNA/culture, and such.

This blog is an attempt to separate the 'wheat from the chaff', recap the bold leaps into AI that somewhat took off in the last couple of years, and show a promise of wider adoption in 2019. There have been some extraordinary forces and visionaries behind these mega AI movements. And then there's talk about the other mega movements that are brewing to become big in 2019.

Autonomous Vehicles

This is perhaps the most interesting use case that is familiar to all of you? GM spent more than half a billion dollars in acquiring a self-driving car startup and recently announced an R&D facility for cruise automation. Ford announced $1B investment in Argo AI, the robotics company created by former Google and Uber leaders. They are planning to launch a fully autonomous vehicle in 2021. Honda plans to include self-driving technology by 2020 via Waymo. Toyota invested over $1B in their R&D to develop robotics and AI technology and launch a fully autonomous vehicle in 2020. Nissan and Volvo have a similar plan to become truly driverless by 2025. Nissan is planning a partnership with Microsoft and Volvo is partnering with Uber to get there.

Many such plans by other car makers, but one company made it a reality with autopilot, electrification, software, a think-outside-the-box design. And then by reaching the market quickly with smart marketing and successfully crossing the chasm by introducing a car-for-the-masses.

Sure, there might have been a lot of backlash around the implications of Tesla and the mind of Elon Musk. But, the fact remains – they changed the auto industry forever. A significant spike in the adoption curve in 2018, especially after the release of the Model 3 in June of 2017 proves the ubiquity of ‘autonomous’ making the Tesla ‘cool’ and ‘desirable’ and a mass market car at a decent price point (Unlike the Model X and S). They did this by competing with luxury cars in a totally different segment.

For those who’ve read Inside the Tornado (after Crossing the Chasm), it almost seemed like the Tesla is in the tornado catering to a mainstream market, but it might be too early to say that considering that it does not comply with all of Moore’s conditions – like demand > supply for example.

Conversational AI

Shifting gears here, in my opinion, conversational AI ranks pretty high on the usability radar, and this particular type of AI definitely crossed the chasm maybe even before 2018. Most of your households have become very familiar with Alexa and Google. Apple and Facebook have also created powerful messaging platforms way before this year.

Forrester’s predictions about brands ‘pouring in’ to utilize the power of these chatbots in their brands in 2018 turned out to be accurate. The smart speaker sales are pegged to reach $28B in 2022. Additionally, AI applications are expected to increase to $37B in 2025, a spike by a factor of 56 since 2016, especially appealing to the Millennials – a powerful consumer segment.

While this has definitely crossed the chasm – 2019 should be the year when chatbots will evolve to listen for intent and emotion some more, not just action. There is talk about how AI will sow the seeds to become EI (emotionally intelligent). Here’s a cool infographic that talks about what’s going to be big with Chatbots in 2019. Although we say this, this is not an easy feat. We might make some progress in this area but the verdict is out to see if we will do any disruptions.

Forrester talks about digital transformation going pragmatic in 2019 with surgical efforts where every company will try to rank up high with points. If you have not already downloaded this report, please do.

Generative AI

Another practical use case of AI that took off this and last year is where the model generates new data that’s very close to the old data. The type of AI drives the next generation of apps with visual programming, creative design, and content development. So, generating new text, music, art, videos photos, etc. and all business models using that fall in this quadrant. Generative AI via generative graphics that can abstract visual patterns and render more ‘intelligent’ versions close to what humans could do.

According to InfoWorld, by 2019 most AI providers will offer these tools and libraries for building AI-powered natural-language generation, image manipulation, and other generative use cases. The e-commerce industry saw a disruption in visual search via this type of generative AI already. Magento has an interesting infographic on AI in e-commerce in 2019. It covers recommendation engines and pattern matching as one of the biggest things that will dominate the year.

We’re now seeing the use of this in emails for Smart Reply – I know a few of you might have used it all it already.

(Cognitive) Robotic Process Automation

This was a hit with the C-level suite especially with CIOs although not all of this is driven by AI – yet. Idea – software takes over the standard tasks with robots doing the work, and with AI it will be ‘cognitive’ bots. Most times these robots interact with the UI of the app which is a challenge where older apps don’t have an API. According to the state of the enterprise automation report from CA Technologies, of 1000 surveyed, 20% use standard bots but want to move to an API driven approach. 

Anyways, I am digressing. Coming back – what does cognitive RPA mean? It is an integration of cognitive capabilities including NLP, ML, and generative AI into the standard tasks. A lot of companies have already shown great promise in this field, and 2019 is going to see a lot more of it.

Some predictions here by Forrester – more than 40% of the enterprises will create state-of-the-art digital workers by fusing AI with RPA. The RPA market will reach $1.7 billion in 2019 and $2.9 billion in 2021. By the end of 2019, automation will eliminate 20% of all service desk interactions, due to a successful combination of cognitive systems, RPA, and various chatbot technologies.

More on this topic in a standalone blog as it’s quite interesting to see the transformation.

 The Road Ahead

So, what more of AI will be hot in 2019? AI models for containerization and orchestration with Kubernetes Clusters. KubeFlow – case in point. Edge computing and Auto ML to create AI models to democratize AI. Supervised learning still time-consuming – so there might be attention towards more reinforcement learning. The generative model will mature a lot more in 2019 and will have more widespread uses cases. A significant progression in hardware with GPUs and quantum devices. Conversational AI will continue to the trek to become emotionally intelligent and autonomous vehicles will go to war. The wheat from the chaff.