The Artificial Intelligence race is on. Companies are vying to patent and exploit its seemingly endless possibilities, with the technology becoming ever more accessible and cheaper.
It’s a myth that AI is the preserve of tech companies and large corporations. AI is already being used by smaller businesses in myriad ways, from automating accounting and HR processes to bolstering the labour force. Indeed, it may help fledgling enterprises take on the big players by lowering barriers to entry.
Moreover, implementing AI doesn’t have to be costly or time-consuming. Many of the productivity tools at its cutting edge are open-source, but the SMEs who stand for gain most from AI are the ones already training their existing workforces in the skills they need to exploit these emerging technologies.
AI jargon defined: what it really means
Artificial Intelligence is, broadly, machines being able to carry out tasks under their own agency. This includes things such as robots and chatbots.
Machine learning is the process whereby a machine learns by sensing whether its decisions are correct and altering its approach accordingly.
Algorithms (the rules of AI) are usually defined as supervised or unsupervised; they’re corrected by an external source (supervised) or left to discover their own pattern (unsupervised).
Natural language processing (NLP) is the process whereby computers analyse, understand and derive meaning from human language in a productive way.
Chatbot . A program that can converse with human users (usually online).
Data mining . The process of sorting through large data sets and identifying patterns and relationships to solve problems.
Open-source software . Software whose source code is open to anyone to customise.
Why AI is big in 2018 (and it’s only going to get bigger)
What is AI?
AI, broadly, is the concept of machines carrying out jobs in a way that we would consider intelligent. Though we might immediately think of robots serving our lunch or washing cars, in practice this covers everything from problem-solving to design and manufacturing via customer service and logistics.
PwC research predicts that by 2030, AI could contribute up to $15.7 trillion to the global economy. This will come mostly from productivity gains as a result of businesses automating their processes – and even increased consumer demand for more personalisation or products enhanced with AI.
Research firm Gartner predicts that by 2020, AI technologies will be pervasive in almost all new software. It also says that by 2021, more than 50 per cent of enterprises will be spending more per year on bots and chatbots than traditional mobile apps. 5G – due to be tested in South Korea in 2018, but not expected to arrive in the UK until around 2021 – will open up huge opportunities for high-bandwidth mobile applications.
The benefits for SMEs are manifold, says Tom Charman, founder of city exploration app Kompas, which uses machine learning to suggest relevant visitor attractions and points of interest: “SMEs are using AI as a low-cost way to get insights from data, thus giving them the ability to make decisions more quickly.
“The combination of both gathering and acting on data, combined with the flatter structure or smaller size of SMEs, gives them the ability to challenge larger companies in a way that may not have been thought possible prior to AI becoming more widely available.”
3 managerial challenges
There will be other challenges for larger companies. A report from Boston Consulting Group and MIT Sloan Management Review points out that success in using AI requires “more than data mastery”. There are managerial challenges in making the transformation to an AI-driven company including:
- the ability to change cultures, processes and procedures
- openness and closer alignment between business and technology strategy
- effective collaboration
Already, manufacturing companies are exploiting AI to do things such as analyse large data sets and determine when maintenance should take place based on the condition of existing equipment. The car industry has been using AI to develop autonomous vehicles and semi-autonomous features for some time, with China publicly testing vehicles since 2015. In fact, all new Tesla vehicles now come with the hardware ready-installed for fully-automated driving, including features such as surround cameras with 360-degree visibility, sensors that can detect hard and soft objects and a radar capable of seeing through heavy weather.
In financial services, AI is leading the way in more personalised financial planning, better fraud detection and anti-money laundering as well as process automation (both back-office and customer facing). Some companies are using AI to model scenarios for capital planning or using natural-language processing – the automatic manipulation of speech and text – and software that interprets graph data for compliance reviews.
AI platform Sentient, for example, creates trading and investment strategies by analysing previously unseen data patterns. AI wearables are being developed to make payments more secure. These will allow fintech companies to better predict how people will spend their money and what advice they need.
How AI can (and is) saving businesses money
The legal sector is undergoing a radical change as AI starts to allow more efficient – and therefore cheaper – ways of working. For example, virtual lawyer Contractor Calculator can power through complex tax law in about 15 minutes.
A report from Deloitte suggests we’ll see fewer traditional lawyers in 10 years, but also a new mix of skills among the elite. It predicts that 39 per cent of jobs (114,000) in the sector are likely to be automated in the longer term. This, of course, has huge implications for smaller businesses who may no longer need to pay for costly legal services, again lowering the barriers to entry.
Technologies such as Beagle, a compliance-management system able to foresee problems and recommend ways to avoid risk, are helping companies streamline decision-making processes. Tetra is an AI bot that converts a phone call to a text-based synopsis of the conversation, including which party said what – as well as having a search feature to recall specific points.
General AI-based productivity tools are becoming more readily available and increasingly affordable – sometimes even free. They include:
- Bot-building platform Chatfuel. The bot can be used to do things such as respond to Facebook queries, post on YouTube and find out more information on potential customers
- Task manager Gluru, which identifies what users need to do from the activity in their email accounts and suggests correct responses
- The Smarter:Time app, which allows users to track their activity – from sleep to fitness to work – to help them use time more productively
- Purple, which provides hospitality companies with insights into customer loyalty via free Wi-Fi services
- Enfocus, a PDF-checking service that can correct errors automatically or alerts the user
- Employee-referral app Woo, which has robotic headhunters
- Conversica, which automates the early stages of sales conversations, identifying strong leads
Apps like these are allowing smaller companies to automate time-consuming tasks and better understand their customers, giving them the chance to compete on a level playing field.
AI in medical insustry SMEs
Healthcare is changing enormously thanks to vast amounts of available patient data. A raft of SMEs are seeking to augment human intelligence with AI, sifting through scans and suggesting treatments based on demographic data. In the future, says Mat Hunter, managing director of the Central Research Laboratory and founder of an accelerator programme for product-based entrepreneurs: “We might see AI offering standalone advice for simple self-treatable conditions, but for now AI is augmenting rather than replacing the humans.”
Scottish SME Aridhia helps transform clinical research into clinical practice using its own technology and a data-analytics platform. The technology is helping biomedical research, precision medicine and healthcare communities work better with people, tools and data in project-management situations.
AI in recruitment
AI is also rapidly transforming the world of recruitment, where unintentional bias is rife . A 2016 study from German independent research institute IZA found that a Muslim migrant woman who wears a headscarf has to make nearly five times as many job applications as a native German candidate with the same qualifications to get an interview.
Xavier Parkhouse-Parker, co-founder of recruitment-tech company PLATO Intelligence, says: “Recruitment is, unfortunately, one of the places in business where unintentional bias dramatically affects the decision-making processes. Heuristics, the instincts that help humans make snap decisions on minimal data, used to promote survival, but now cause biases.”
AI using the right data can eliminate almost all human bias. By means of algorithms, the technology is able to make accurate predictions about the right candidates in a way the brain simply can’t compete with.
Weeding out unwanted or irrelevant applications, sifting through CVs to find the right information, booking rooms, cancelling and rescheduling appointments; recruitment all takes time. The industry is using AI to understand key information and make appropriate recommendations within minutes, allowing humans to focus on the interview.
One of the most difficult things a recruiter or HR manager can face is good staff leaving. But AI is already able to make strong predictions about which candidates are likely to quit. So that all-singing, all-dancing applicant might exhibit some behaviours and language that suggests they won’t stick around.
How to build trust in Artificial Intelligence
How people react to AI is crucial. It might be assumed that customers, job applicants and staff would be cautious, however most systems have been created to build trust.
“AI is an efficient method of making decisions based on huge amounts of information. For this reason, many people consider AI to be an invasion of their privacy,” says John Regan, CEO of behavioural insight company Human Technology.
He adds: “The reality is that although the information processed is much greater than in the past, it is automatically processed by a machine. Any bias or prejudice is proactively sought and removed from machine learning, and the way in which information is held and processed is unintelligible to humans.”
This is a key takeaway. Businesses need to educate users about how data is being utilised to personalise information or make decisions, says Tom Charman of Kompas: “To create trust, users must feel that AI is acting in a way that benefits them, as opposed to the business or organisation implementing the technology.”
Charman says a set of rules is essential – such as the 23 Asilomar AI Principles – to ensure AI remains a force for good. That, he believes, would lead to a system that not only regulates AI, but builds a level of trust between businesses and the end-user.
Dystopia? Not really. AI will benefit enterprise
At the Future of Life Institute's Beneficial AI conference in early 2017, Google DeepMind CEO Demis Hassabis predicted that superintelligence would be developed just a matter of years after machines achieve human levels of intelligence.
The conference discussed whether investment in AI should be accompanied by funding for research into ensuring its beneficial use and focused on issues such as how we can get future AI systems robust to do what we want without malfunctioning or getting hacked. It concluded that advanced AI could represent bring profound change to our way of life and must be planned for and managed carefully.
Change happens fast in AI, says Regan. “Amazon first announced its intention to develop autonomous delivery drones in December 2013.” At the time, he says, “the announcement was considered by many to be little more than a PR stunt. Four years later, it’s commonplace.” Now drones that take selfies, return to their takeoff location when they run low on battery and avoid obstacles are on sale in the UK for less than £100.
Researchers predict that among the next big developments will be better language understanding. Regan says: “AI-driven PAs such as Siri and Alexa will reach the point where they are able to process plain English requests and understand context and nuance. More effective voice control combined with augmented reality will also enable more physical and intuitive ways of interacting with computers. The result will be that computers will be less obvious, but increasingly ubiquitous.”
AI is beneficial, disruptive and transformational. Most importantly, it is here to stay. So it’s essential that businesses large and small are equipped to make the most of it.
Prepare your business for AI in 4 steps
There are many reasons to be optimistic about ai. an accenture report suggests that the technology could bring an additional $814 billion to the uk economy by 2035, while japan has the potential to more than triple its annual rate of gross value added (gva). ai will change how companies work, large or small. gartner says that in 2020, ai will create 2.3m jobs while eliminating only 1.8m.
Step 1: Try out small amounts of ai
don’t invest too much or too little. tom charman of kompas says, “there are already lots of off-the-shelf tools that can be used quickly to begin to create analysis on data you already have and won’t require the same level of skill compared to building the technology from the ground up.”
Step 2: Train your workforce.
Once you’ve achieved some basic implementation of the technology, it’s a case of making sure you have the right people to continue to develop it and implement new processes. “the way to do that and avoid being left behind is by looking to data scientists,” charman says. “if your processes require more skill, look to machine-learning researchers. these are people who will quickly be able to analyse and interpret data sets, tweaking and refining where necessary to improve efficiency.”
Step 3: Consider partnering with an ai software business.
John Regan says of human technology, “there are many start-ups developing solutions based around machine learning and ai. it is likely that you will be able to identify a business working on a solution similar to the opportunity you’ve identified within your own. most businesses in this category are keen to work with potential clients to tailor their solution to suit a particular requirement as this gives them an excellent opportunity to research and develop their product.” this is a lower-risk opportunity since such businesses often take on the responsibility of project management as well as providing the expertise.
Step 4:plan for the future.
Understanding what human staff you will need is crucial. consider what processes could be automated. says futurologist philip ross, “organisations need to give their attention to curating talent or being more flexible to people’s work preferences to retain the best.” he adds, “ai needs to be introduced into the workforce more rapidly. this allows organisations to focus on the jobs which require specific skills and find the right talent to fit those positions.”