AI Tools and Applications
AI can be embedded into a range of digital tools, such as:
Chatbots – often used with messaging apps to guide or direct a conversation towards an end goal, such as a sale or a service.
Search – search algorithms use AI to analyse the words and phrases that make up a search query.
Analytics – AI can be used to obtain hindsight, insight and foresight.
Risk management – AI helps deliver higher accuracy in investigations and detections and reduces false positives.
Documents and contracts - more efficiently, accurately, and cost effectively extract and summarize key provisions from legal documents.
Voice processing – use voice as an input into a computer, or process voice data to extract and process meaning.
Image processing – use images of video as input into a computer, or extract meaning from them.
IT Service Management (ITSM) - improve IT service delivery with enhanced monitoring and service desk functions.
Generative Design – speed up and improve the design process using goal directed design (GDD).
Embedded – driverless vehicles, shop-floor robots.
AI Applications
We can now train computers to perform better than humans in many tasks. Take CT scan analysis for example. One person would take over a thousand years to examine 30 million scans, whereas an AI system called Enlitic can perform the same task in just over a week, and to a higher degree of accuracy[i].
Financial gains from AI can also be direct and quantifiable. For example, Emma, a chatbot at OCBC bank in Singapore, turns 10% of chats turn into loans. It uses Natural Language Processing and personality analysis, and the result is that customers feel “this interface understands me”. In its first eight months of operation it won $70m of home loan business.
Careers which were previously thought safe from automation are also being disrupted by AI Take legal practice, for instance. UK Law AI company, CaseCrunch, pitted lawyers from many of London's top firms against an artificial intelligence program called Case Cruncher Alpha. Both the humans and the AI were given the basic facts of hundreds of claim-based cases and were asked to predict whether a claim would be granted or not. Case Cruncher got an accuracy rate of 86.6%, compared with 66.3% for the lawyers[ii].
[i] https://www.accenture.com/gb-en/blogs/blogs-could-AI-become-doctors-secret-weapon
[ii] http://www.bbc.co.uk/news/technology-41829534
Figure 10. AI is increasingly used in the practice of law. Image - LawGeex
Financial Service is one industry that is pushing ahead quickly with the adoption of AI Applications in this industry include:
· Security - monitor & intervene, including Dark Web crawling
· ‘Know Your Customer’ (KYC), identifying and verifying the identity of clients
· Face recognition - pay with a smile, or, as used in China, recognise if an applicant is not telling the truth
· Personality analysis – ‘we know by your traits’
· Sentiment analysis - we know how you feel’
· Predicting loyalty
· Sector/geography/company analysis
· Algorithmic trading using ‘duelling’ algorithms
Given that there are so many potential applications of AI, it helps to categorise AI applications in relation to the benefits being sought. Organisations will require different categories of solutions depending on whether they need AI for efficiency, effectiveness, innovation or expert systems. The kinds of data needed, and the deployment options will also be different for each scenario.
Figure 11. Categories of applications of AI