AI has continued to rapidly advance. With intelligent tools now available to businesses, how can enterprises transform their processes and insights using the latest AI techniques?
If your business doesn’t have an AI roadmap, it will suffer significant losses and damage your ability to compete. This is the stark message that many industry watchers are defining, as we enter a new era of AI.
The headlines in the technology press that shout that AI is the next seismic shift in how your business will operate, need to be tempered with a reality check. AI technologies as they stand today can deliver real, tangible benefits in a number of business areas most notably HR, customer services and marketing, but AI is not a technical panacea.
“The past 18 months have been an exciting period for researchers in natural language processing,” Ben Lorica, Chief Data Scientist at O’Reilly Media told Silicon: “These days the favoured approach is to use an off-the-shelf model, pre-trained on massive amounts of data. We are going to start seeing these developments used in commercial products. Researchers and practitioners are beginning to tune these models for specific application domains. There are also promising results that these pre-trained models can be used to create knowledge bases.”
With Professor Giovanni Miragliotta, Professor of Advanced Supply Chain Planning at MIP Politecnico di Milano Graduate School of Business – whose research has covered IoT, cloud computing and AI, also commenting: “It is hard to measure how businesses have taken into consideration AI capabilities both past and present. However, I think we may agree on three things. Firstly, avoid over-expectation when it comes to incorporating AI into your business structure. Start from the bottom and don’t plan beyond your organization capabilities. Secondly, it is a long journey and companies should plan time and money not just to start up a project, but also to train algorithms correctly during the time. And thirdly, involve internal users when starting new AI project, to avoid long term organizational issues.”
AI is used as the general umbrella term for machine learning (ML), deep learning and natural language processing. AI tools had to wait until computer processing power was sufficient to make these services, something businesses could buy and implement. Coupled with the arrival of cloud services, the current state of technological advancement has allowed AI to become a practical reality for your business.
Says Deloitte in their report into the impact of AI on business: “AI modelling and simulation techniques are already being used to help drive sales in several ways. Platforms providing real-time data gathering, forecasting, and trend analysis can offer greater insight into buyer behaviour. Artificial intelligence used in conjunction with CRM systems can automate functions such as contact management, data recording and analyses, and lead ranking, while AI-enabled buyer modelling can provide a prediction of a customer’s lifetime value. Recommendation systems are also used to learn content preferences and push content that fit those preferences, allowing more targeted sales.”
In essence, AI is about how the masses of data your business contains can be utilised. Today, the most successful companies are those using the data they have to deliver effective changes to their enterprises. These changes cover a broad spectrum from enhanced customer services to new product and service development. AI now plays a critical role in how your business will evolve into the future.
Until recently, AI was confined to specific tasks such as beating chess grandmasters. Today the technology these achievements were based upon have evolved into AI as we know it today.
AI is now having significant impacts on several key business processes: Automation that uses Machine Learning (ML) is driving a revolution in how businesses present their customer services. Image recognition is being used across the manufacturing space, and as a critical component of security systems on digital devices.
O’Reilly Media’s Ben Lorica said: “Computer vision is probably the area with the most activity, measured in terms of patents, number of start-ups and use cases. Deep learning has proven to be adept at several perception tasks involving images and video. The applications span across security, logistics, manufacturing, healthcare and medicine, media and advertising, agriculture and many more.”
Lorica continued: “As well as this, speech technologies have proven beneficial for all, particularly as the amount of audio data explodes. It is estimated that by 2020, about half of all online searches will use voice. Smart speakers are also projected to grow by more than 82% from 2018 to 2019, and by the end of the year, the installed base for such devices will exceed 200 million.”
Some of the most exciting applications of AI are in predictive analytics. Here, the masses of data now collected by your business can be interrogated to reveal new patterns that can help your company predict the online behaviour of your customers, for instance. AI is excellent at pattern recognition, which is being applied to combating payment fraud.
However, you can mine the information contained in your business that is the most transformative aspect of AI. For example, Walmart makes excellent use of SAP’s HANA for predictive analytics, with Avanade (developed by Microsoft and Accenture based on the Cortana Intelligence Suite), is driving predictive analytics, as a tool any business can buy today. Business and IT leaders predict they will see an increase of more than one-third (33%) in revenues from the use of smart technologies over the next five years.
The core use of AI is to enable your company to make better decisions. The investment you make today in these systems and technologies will give your enterprise a foundation to build onto. Business intelligence must have an AI component. As the quantities of data, your business collects now far outstrips the ability of your staff to meaningfully process it, automating some of these tasks is a priority.
Saryu Nayyar, CEO of Gurucul – a global cybersecurity company, commented: “Many innovative organisations have used AI very strategically to accomplish specific business strategies. For example, one large American furniture company used the most common Google searches for furniture, like a red chair, to create specific domains for websites. They did this for more than 250 search terms and managed to outmanoeuvre companies like Amazon. There are other examples like this of organisations that were able to innovate with AI and this will become more common as AI becomes more pervasive.”
An AI future
Care, though, must be taken when developing AI systems for your business.
An essential component of using AI is to do so with clearly defined ethics and explainability. Your business must ensure its use of AI systems is robust and doesn’t have a bias. Also, you need to ensure your AI system’s output, can be explained primarily to regulators. Ask yourself, why your AI system flagged some data or customer behaviour (perhaps, as part of your fraud prevention), and, understand if these are actually false positives. It’s vital to maintain customer trust that your AI systems are working accurately.
Dr Anthony Scriffignano, Senior Vice President and Chief Data Scientist for Dun & Bradstreet, commented: “There are many types of bias. Some of the major ones are confirmation bias (going into a problem with pre-conceived conclusions) and convenience sampling (using data you have on hand, instead of a more representative or appropriate sample).”
With Gurucul’s Saryu Nayyar also explaining to Silicon: “Businesses should keep in mind that AI is not magic which can be used as a cure-all, despite the claims of some vendors who utilise AI technology. Businesses just need to make sure they avoid the marketing hype that’s sprung up around AI. They should start by identifying the specific cases within their businesses where AI might help. Have a particular objective of how AI might help improve processes. Also, remember that the real promise of AI comes from the data that feeds it. To be successful, you need to ensure that you have the appropriate data. If you don’t, the output could be flawed and, you’ll go down the wrong path.”
To successfully implement an AI strategy across your business will require a detailed understanding of the data your company has, and what insights you want your AI to deliver. AI tools are available now. Placing them within a clearly defined strategy will always result in better outcomes.
Paul Lasserre, Vice President of Product Management for Artificial Intelligence, Genesys concluded: “AI is no longer something that will impact businesses in the distant future. It is here now, and companies are already taking advantage of its ability to produce targeted outcomes, assist employees and tackle old problems more efficiently. Today, AI is more accessible than ever; it is no longer reserved for large enterprises and smaller businesses are also using it across several different functions. By leveraging the cloud’s ability to process enormous amounts of data quickly, AI-applications are being used in nearly every aspect of the business today, such as marketing, sales, customer experience and administration.”
Your business’s future will include AI technology. As the cloud has transformed other areas of your company, AI will change how you approach data analysis and HR in particular. We are, though, in the early stages of this technology’s development. Take your time to assess the tools available and whether they can bring tangible benefits to your company. There are few components of your business’s operations AI won’t touch.
Silicon In Focus
Peter van der Putten, Assistant Professor at Leiden University & Director AI Solutions at Pegasystems.
What can businesses learn from the history of how AI and how this has been used within enterprises?
“It is natural for business owners and managers to want to rush into the adoption of technology innovations. It happened first with mobile twenty years ago, again with the shift to social media channels in the last decade and most recently with AI in CRM (customer relationship management).
“It’s tempting for businesses to do the corporate equivalent of sticking their fingers in too many pies – deploying an offering just as soon as it’s made available, and perhaps before the tech has had time to mature. More often than not, this strategy results in multiple disconnected systems for each solution which are spread across the enterprise. Other than being difficult to manage, the inconsistent customer experience this creates will be more frustrating to customers than the technology is beneficial to businesses.
“Instead, businesses should take a holistic strategy to adopt AI, which places an ‘always on customer’s brain’ at the heart of the enterprise. Connecting customer-facing channels to this brain is the best way of leveraging the full power of your data, making sure you are always on the same page as your customer.”
Can you outline some great current uses of AI? Who is using to the best effect and why?
“AI in CRM enables larger enterprises to optimize customer experience and value, one interaction at a time. Royal Bank of Scotland manages 17 million customers but managed to raise its Net Promotor Score by 18 points unanimously. Sprint’s adoption of AI in the cloud achieved a 14% increase in customer retention in just six months, simultaneously overcoming an industry-high in turnover rates.
“In Australia, Commbank uses an AI-driven Customer Engagement Engine to drive meaningful conversations with their customers across all their channels, from mobile apps to branches. This goes beyond just pitching the next best product to sell. For example, they alert customers who are about to incur fees, advise them on specific government plans they may qualify for based on what the bank knows about them, or switches to Next Best Conversations related to wildfires for customers in affected areas.
“Businesses which are adopting AI strategically to address existing problems are the ones who are truly harnessing its benefit. Although AGI (artificial general intelligence) is still a long way away, platforms which can serve as the central AI brain connecting CRM applications ranging a sizeable enterprise has redefined how businesses engage with their customers.”
Which key areas of businesses are being impacted by AI?
“Today, AI is at the heart of some fundamental processes that affect profit and loss in meaningful ways. For a bank, machine learning in its marketing systems continually improves what recommendations to deliver to a customer by learning from customer feedback in real-time, quadrupling sales effectiveness. For an insurer, robotic process automation sequences the activation of apps and data to cut out dull, mundane tasks and speed up the claims process. By using text analytics on inbound email, a bank can automatically route customer service emails to the best team for 70% of the incoming emails.
“Perhaps not as exciting as human androids, but these are examples of AI and software robotics embedded in everyday business processes. For example, robotic process automation does a massive job in helping to augment the ability of their human co-workers to be more helpful and supportive with customers.
“AI technology can be used to streamline and simplify business processes across different siloes, which is key to reducing costs and improving customer experience and value. For example, intelligent software can facilitate access to the right data, which makes it easier for client-facing personnel to meet customer needs. Far too many businesses have fragmented systems with disconnects between their front and back offices, which is a major obstacle to effective customer service.”
What are the risks and pitfalls businesses need to avoid as they use AI across their enterprises?
“As we enter the age of responsible AI, eliminating unwanted biases which exist in data, and providing transparency and automated explanation facilities will be a priority. For example, are models blindly discriminating against groups of certain gender or age, or is there a bias in how data for these models was collected – for example, was fraud only investigated for certain groups? Instances of suspicious outcomes in AI has opened up the discussion of whether transparent methods should be preferable to opaque methods such as deep learning, for ‘explainability’.
“Emphasis on the design and development stage of AI modelling when it goes wrong has been another factor under scrutiny. The focus is now on building models which can bring insight into the customer interaction to have a positive impact on customer experience. Harder still is operationalising this insight. The deployment, monitoring and constant fine-tuning of AI assets such as predictive models is the most challenging problem for most organisations, as well as how to combine state of the art machine learning with business rules, strategies and policies from ‘Good Old Fashioned AI, to keep models under control and translate predictions into actual automated decisions.”
What’s the future of AI look like?
“The future of AI in CRM is not one that is channel-specific, but one that is channel-less in nature. With an AI brain at the centre of your customer experience platform, businesses can react in real-time with the next best action regardless of where customers are on their brand journey or how they engage – be it a chatbot, mobile app, web page, on the phone, or in-person at the store.
“This singular AI brain approach allows businesses to extend predictive intelligence to all other channels, without having to start from scratch for each new interface that comes along. Business users will always require control, if not for transparency, then to implement corporate strategy and policy. So, it will be a future where augmented intelligence will be key, not just artificial intelligence. Rather than replacing man, we foresee a more mutually beneficial symbiotic relationship between man and machine.”