If the buzzword for 2018 was Blockchain, for 2019 it is AI implementation.
Every day, business blogs, news articles, and think-pieces talk about the coming of AI. From its risk and dangers, to its potential to transform the way we all live and do business.
These are all worthwhile things to consider and be aware of. However, most of these articles concentrate on the next 10-50 years or so.
What is far less talked about, is AI implementation today. How can AI make a difference to you and your business this year, rather than in the next twenty?
To answer this, we first have to go back in time…..
Imagine it’s 1972 and you’ve just invented email.
This is an electronic version of regular mail which is delivered almost instantly. It uses no paper, no ink, and is far more reliable. All it requires is a computer and an internet connection – and it can be delivered anywhere. You can clearly see how this technology will change the world.
However, to the people around the world using typewriters and regular mail, it’s not clear how it could immediately help them.
After all, you can only email someone who has an email address set up – which means this person also needs a computer and an internet connection. For it to be useful within your company, you need to make sure everyone now has a computer. If you want to talk to other companies, or people globally, they too need lots of computers with internet.
This analogy shows some of the problems with implementing AI. Sure, self-driving cars sound wonderful. If all cars were autonomous, a centralised computer with a birds-eye view of the entire city could manage traffic efficiently and safely. But they’re much harder to use and train when surrounded by irrational, human drivers. Benefits can only be had when the correct infrastructure is in place.
In the case of email, it only became widely used, and entered the general lexicon, from 1993 onwards – over two decades after it was invented.
Casting our eye over other vital technologies, we can see that the same principle applies. Much of the time, you need to install entirely new infrastructure to reap the benefits of a technology. Think of the national grid, our freshwater system or gas pipes. These all work wonderfully, but only because the infrastructure has already been built. Having a home with radiators and taps is no good if you can’t connect those pipes to anything.
So how does this relate to AI implementation?
AI is an extremely broad term. It encompasses things like machine learning, OCR capabilities and algorithmic optimisation. Ultimately, as with all technology, it needs to be applied to fix a specific problem. This could be to intelligently automate a process, or to assist in a complicated long-term process.
The fact remains that AI, much like every technology, requires some infrastructure to operate. To reap its benefits, what will we have to change?
Similarly to electricity needing gridlines, and email the internet, AI requires structured data. That is to say, it needs a pool of resources to learn from, datasets by which to recognise what solutions are ideal, and what solutions are less so.
Systems built in the last decade weren’t anticipating the arrival of AI. For most organisations, their data is scattered across disparate systems. Legacy estates of servers, siloed software tools for specific functions and so on. Step 1 of implementing AI today would be to get this data in a structured format.
One way to achieve this is an elaborate data migration to an entirely new infrastructure. This is costly, complicated, and prone to error.
Another way, which we at psKINETIC far prefer, is instead to integrate existing systems into an Appian Platform.
The Appian platform is a highly configurable digital transformation platform. It can integrate into virtually any system. So, whether you have 23 disparate systems or a coherent server-software setup, Appian can function as a top-level ‘layer’. It unifies your data into a structured record.
Once data is structured into records, it becomes exceptionally easy to leverage AI in its many forms to your advantage. Whether it uses Google OCR or some other machine learning, Appian has custom-built integrations with these services. So, as AI tools become more sophisticated, you and your organisation will continually have easy access to the latest and greatest.
Ultimately, Appian prepares your business for AI implementation not only tomorrow, but today.
psKINETIC, utilising AI applications such as Google OCR and sentiment analysis, have already started leveraging great value from AI for our clients. If you’re interested in learning more, please contact Vikki Hailey or visit the AI Solutions page on our website.
AUTHOR: Jack Lawrence | Digital Consultant | psKINETIC