Artificial Intelligence is changing the way processes are carried out across a number of different sectors. As the technology becomes more accessible and affordable, what does it mean for organisations in the SME/MLE sectors.
In 2016, AI has become one of the most disruptive forces across a number of different business sectors and is revolutionising the way various processes are being carried out. But what is AI, how does it work and how can it help Irish businesses?
What is AI?
AI refers to the creation of intelligent machines/computers that can self-learn and even go on to think for themselves, but not necessarily in ways that we will always understand.
What has this got to do with business? Well, it all starts with ‘big data’. Thanks to tried and trusted data collection practices, that have only been enhanced further by advancements in technology, in house data scientists and analysts have more information on their hands than ever before. This data is analysed, captured, curated, searched, shared, stored, transferred, visualised, queried and updated, and the learning outcomes are used to help companies make informed business decisions.
Now imagine that instead of a human doing all the work, you had a machine, with a fancy algorithm that was being given access to a vast amount of data sets. The more information that it is fed over time, the more accurate the predications it can make, at faster speeds. It learns from its mistakes and is constantly setting up rules and exceptions so as not to repeat them.
So, simply put, AI is all about behaviour, context and prediction. It’s how a machine makes a decision based on the information it has available to it.
AI in the Business World
Bots
One segment of AI that is bound to be of interest to any companies involved with B2C communication is ‘bots’. Bots are a software application that run simple and usually repetitive tasks online. If you have ever used SIRI, the personal assistant built into the iPhone, you may be used to asking it to do simple processes like mark an appointment in your calendar, tell you what the capital of a foreign country is, or do some calculations for you.
While SIRI, for the most part is voice activated, a lot of bots take the form of ‘chatbots’. These are bots that live in messaging apps and respond to a text input from a user. In the past they would have required a request written in an exact match structure, the same way writing commands in computer code would trigger an event. Now these bots are being programmed to recognise, respond and take action to queries through a conversational medium.
A great example of bots in action is within Facebook Messenger. Recent updates to the social media messenger allows brands to reach out to customers with generic message templates that can market promotions and special offers. For example, a clothing retailer might send a form to a customer that has opted in to its service that advertises a style of shoe it might like and give three options: “Buy now, see more items like this, or contact customer service”.
For developers looking to design their own bots and go far beyond generic templates, building with bots gets even more exciting with the wit.ai Bot Engine.
This allows companies and their developers to continually train bots by using sample conversations and enables businesses to create conversational bots that can automatically chat with users. The wit.ai Bot Engine effectively turns natural language into structured data as a simple way to manage context.
In other words, over time bots for a travel agency would be able to deal with customer queries that were posed in natural language. So instead of a user typing “Hotels, New York,” in a search engine in order to bring up some onscreen options, they would pose the request to the company’s bot via Facebook Messenger in casual conversation “Can I book a room tonight?”. The bot recognises the request, puts it in context of time and date and searches the company’s reservation system before responding with a personalised answer.
Financial Trading
Considering the millions of trades that take place every minute in various financial markets all over the world, it’s no wonder that intelligent automation would find a home in the finance sector.
In an article on Raconteur.net, it was revealed that algorithmic trading systems now handle 75 per cent of the volume of global trades worldwide and this figure is predicted, by those in the industry, to continue to grow steadily.
Pension managers, stockbrokers and others working in the sector can benefit especially from algorithms focusing on risk management. These are able to simulate thousands of risk scenarios within a second, similar to FICO credit scoring that banks use to approve customers for credit products like loans and credit cards. Those using AI for trading are very interested in making sure the machines can learn from past mistakes as a result of ongoing learning, before going on to make even more accurate predictions.
Google Cloud Platform
With the help of the Google Cloud Platform, more and more businesses have access to machine learning. To work with this offering from Google, a business gives the platform access to its system and analytical data so it can generate prediction models from training or base data that the organisation feeds it.
According to Google, with enough learning, you can create applications to perform tasks such as predicting what movies or products a user might like, categorise emails as spam or non-spam, assessing whether online comments have positive or negative sentiment, or even guessing how much a user might spend on an e-commerce site in a given day.
Businesses that use Google AdWords to advertise their products or services can feed the results of previous ad campaigns into the platform and allow the machine to predict the strongest set of components that could generate conversions. In this scenario the platform would look at things like character count in ad copy, position of ad, call to action and any other data it is fed to make an informed prediction. The more data from campaigns that it is fed, the more accurate the prediction could be.
IBM Watson
Watson from IBM is possibly the AI service that most people are familiar with. In its TV ads, Watson is portrayed by a helpful talking black box that, with the help of some celebrities and well known organisations, solves some complex problems. Watson equips businesses with four main features:
- A virtual agent that helps businesses provide their customers with personalised support.
- Watson Explorer which combines cloud-based enterprise search and content analytics with cognitive capabilities to give the clearest overall picture of business performance.
- A company analyser that is a cross between a CRM and a competition analysis tool.
- A knowledge studio that teaches the Watson API specific knowledge from industries of your choice from leading experts.
What Does AI Mean for The Jobs Market?
Much attention has been given to a recent report by Forrester that suggests that by 2025, 7% of the US workforce will be replaced by advanced automation services that are underpinned by machine learning.
The report predicts that AI solutions in the future will be equipped to process and respond to countless complex scenarios that will allow it to handle the daily occurrences that pop up in industries such as transport, customer services and logistical roles.
It’s easy to see advanced bots one day being sophisticated enough to handle all aspects of customer service, but this prediction works under the assumption that AI services will have a much greater comprehension of spoken human language by 2025.
For other industries such as transport, the rise of the machines mightn’t be so straight forward. For example, to make a dent in the automotive industry with the likes of self driving cars, the risk engines would have to be refined in such a level that human safety is guaranteed both for people in the car, pedestrians and other drivers.
What Happens Next?
The computational power of processors will get more powerful, in accordance with Moore’s Law, AI applications will be ubiquitous and computers will go on to think for themselves. Now, this could be some time away yet, but as the world changes, make sure your business changes with it. In the coming years, having an AI strategy could be as important to your business as current infrastructure fundamentals such as business broadband and voice and data solutions.
If you are working in research, finance, or any consumer facing e-commerce business talk to your developers, researchers and any other relevant stakeholders and do the following:
– Identify any processes that could be automated at present or in the future within your organisation that would increase efficiency.
– Identify any AI products or services that are within your operating budget and setup a meeting with a solutions provider.
– Organise your data, whether it’s research data, customer service data or store inventory lists and consult with the provider and stakeholders on the best way to prepare it to work with AI.
– Don’t go ‘all in’ straight away. Integrate AI slowly into one existing area of your business’s operations first and then scale up as appropriate. After all, it’s not just the machines that have a learning curve ahead of them.
Get Your 60% Discount for UXDX 2016
Virgin Media Business is proudly sponsoring UXDX 2016, taking place in the RDS on 2nd November. The conference will feature experts in design, IoT and UX from companies such as IBM, Intel and BBC to name just a few. The Virgin Media Business network can now avail of a 60% discount! To register, go to www.uxdxconf.com and enter the discount code “virginNetwork” on the ‘book now’ page.