Understanding the Tech impact at Google IO 2018

Google IO is the company’s lead tech expo and demonstration event. The three day affair highlights some of the work they are doing and some of the development they are undertaking with their partners. Importantly it shows the direction those tools and technologies are taking. The concepts and ideas presented at the conference often go on to change the tech landscape and with the Google juggernaut behind the ideas they are either very successful or burn out quickly (Fushcia anywhere?).

Consider Android is now 10 years old and installed on billions of devices worldwide and has drastically changed the technology landscape. It was announced at Google IO and now accounts for around a third of the sessions here.

The innovations here can change companies and navigating that landscape requires the right investment in people and technology. I’ll highlight some of the concepts here that I think might make an impact on existing businesses. Adopting some of these technologies can make a huge impact within organizations. (Hey, that’s what I help with)

AI

Given the bandwidth taken up by AI articles and sessions, it is pretty much assumed that it will dominate technology changes in the next 5–10 years. Sundar Pichai, CEO of Google called it an “inflection point” in computing. Google, Apple, Facebook, and virtually any organization with R&D is throwing a huge amount of investment in order to build out the tools.

This is nothing new. For the last couple of years the promise of AI was that it could be used with your existing data and systems. While that wasn’t a lie, it did require you to do a significant amount of programming and data conversion in order to implement the system. Notably the last mile AI, producing actionable results, was always up to the user to implement from scratch. This had the side effect that if you wanted to start with data analytics and AI you were required to jump in all the way and dedicate to the project. This technology investment for existing companies was a risky venture. However, this year introduced a number of tools that can make that process much much easier, cheaper and safer.

APIs

Google and Firebase (Google acquisition 2014) both announced new and revised APIs for development. While they are numerous and expansive, these APIs allowed not only access to their AI tools (vision/text) but the framework to host and run your own models. Those APIs are a quick and cheap way to leverage the skills and investment Google puts into AI and integrate it within your projects.

I was particularly excited around the announced capabilities in text-to-speech. Text/speech and conversational APIs can make a big difference in accessibility and the breadth of your applications. While booking a haircut is a great keynote demonstration, the capability of text and speech APIs could allow companies to automate legacy systems and processes without heavy infrastructure and systems changes.

Data entry can now be handled with voice input. Systems that were traditionally handled with phone calls such as dispatching can be automated without changing workflows. Field management can integrate this technology and link with existing legacy systems without having to rebuild processes or systems. There is a lot of potential to integrate these tools for a much more nuanced user experience and fit in with existing systems. It’s cheaper and easier than rebuilding an entire process and can produce real benefits in a very short time. Automate the parts of your business without requiring a huge investments.

Hosted Models

Second and a little more technical is the ability to host your own AI data models synchronized and distributed across platforms. AI models are the heart of the prediction and the real power of any system. Google’s Vision API is a wrapper around their vision model. But having an API really limits the speed and offline capability of your predictions. Mobile devices have the capability to run these models but managing those models meant you had to create either a download and update infrastructure or bundle it with your app. Changes to the model were difficult to manage and slow to update.

Firebase, now allows you to rapidly synchronize those models across deployments. Consider a theoretical system to identify plant diseases in the field. You build a model with existing pictures and distribute this with a mobile application. However, as new diseases are updated and the prediction model improved you have to distribute this to all the devices. Using the distributed model you can dynamically update the model without needing to update the device. This opens a huge potential for mobile devices to act as both data gathering and prediction tools.

Those are two small examples where that technology investments can impact non-techy businesses. The tools here simplify automation and can make changes in many industries.

Look forward to additional articles on Design, Mobile and Automation in the next couple of days. And if you have any feedback feel free to reach out at john@jotdigital.com