Responsible AI is coming and, like governance and green vegetables, it’s good for you – and for your business.
Get this page as a downloadable PDF by filling in the form below:
Table of Contents
First, a question of grammar: well vs. good
It’s easy to measure if your business is doing well.
Take key performance indicators: profit, number of sales, average transaction value, number of new hires, or any factor that denominates ‘growth’. Is the graph line showing a trend that is upwards and to the right? At that point, most business owners would call it a day. There, they’d say. We’re a successful business. Job done.
If you want to go beyond the bottom line, you need to ask: how can I tell if my business is doing good?
And, what tools or processes can I implement to become a good business, not just a profitable one (though that matters, too)? We’re not here to tell you which mindfulness app to download. What we will explore is how to use AI in key areas within your organisation, responsibly, and how leveraging artificial intelligence is just good business.
‘Most algorithms produce an outcome or answer that humans rely on for making a decision. Most algorithms, however, do not explain how they arrived at that answer. Lack of explainability weakens accountability and does not foster trust.’ – Dr Shefaly Yogendra, COO at Ditto AI
Trust in BPM decision-making is as vital as the decisions themselves. Yes, AI can help to make better, data-informed decisions, but transparency matters. To be a good business, you need to show your workings.
This study on decision-making authorities says that ‘employees use trust-based expectations of how management will treat them as a lens through which to interpret or respond to their actual treatment.’ Without trust, people are less likely to view ANY decisions favourably, whether those decisions benefit them or not. So, any use of AI within the decision-making process must be accountable and auditable.
Repetitive tasks are there to be automated. In any business, this could affect areas such as:
- Data mining
- Social media
- Customer responses
How is this good? It frees time for your employees to tackle more complex brainwork, such as coming up with innovative new business processes or giving a customer with a complicated issue more facetime. As we say in our blog about AI myths, AI and automation is actually predicted to create jobs. In 2017, Gartner found that though 1.8 million jobs would be lost to AI, 2.3 million would be created by 2020.
So, to use AI for good in your business, you need to provide opportunities for your employees to do better, smarter work. Set aside time and attention for management-led training, encouragement, and modern goal-setting.
Instead of having your legal team, accountants and business decision-makers spend hours pouring over the end-of-year taxes, why not use an AI tool? H&R Block and PwC are already doing it. This will save time and, with current capabilities, will probably be more accurate. Deep learning tools perform advanced functions, such as fraud detection, which require some nonlinear reasoning. That is, a way of thinking that doesn’t necessarily have a clear set of steps, but various options that need to be interpreted.
Tax represents your social contract to give back to the infrastructure that helped you build a successful business. It’s worth doing right.
Current news makes it seem that businesses are only just waking up to their obligation to grow a conscience and address the climate change issue. Truthfully, climate action has always and should always be a core concern for responsible organisations. Just 100 corporations are responsible for over 71 percent of emissions worldwide. Kitting your office out with low-wattage bulbs isn’t going to cut it. Business standards need to change.
Thankfully, companies like Microsoft are pioneering ‘AI for the Earth’ and other AI initiatives to counteract climate change. Any business can use predictive AI to parse data and extract key areas for waste reduction, carbon offsetting or other efficiencies.
Over half of HR specialists surveyed by LinkedIn identified screening candidates in a large applicant pool as the hardest part of their job. Using AI powered recruiting software, such as Clearfit, and automating this part of the recruitment workflow would:
- Enhance a business’s ability to find the right talent for the job;
- Lower the chance of bias and provide standardised metrics for better job matching;
- Reduce wait-times and get applicants into their new job quickly. (We’re all familiar with the agonising wait for a call from a potential employer.)
Further work in this area is underway, with chatbot recruiters, digitally monitored interviews (assessing speech patterns and facial expressions), analytical ‘games’ and virtual HR assistants. Not every AI tool is equal. Amazon scrapped their AI-driven recruitment process in 2018 when they discovered it was imbued with a bias against women. Part of using AI well is researching the quality of the provider and the accountability of their product, and testing it before you roll it out company-wide.
Honeywell, a multinational engineering company, are combining augmented reality and AI to reduce ‘information leak’ due to churn, as well as provide training to new generations of workers. They’ve seen improved training times and a better match between what needs to be learned and preferred ways of learning.
Digital factories, the digitised twin of a factory location’s counterpart, can also identify individual training needs with remarkable granularity. Using product line engineering (PLE), based on real-time troubleshooting and predictive analytics, you can see where there are opportunities for an individual or team of workers to improve.
Employee wellbeing and monitoring
Computers have made workers many times more productive and AI promises to continue that trend. With AI, everyone gets superpowers that connect our global intelligence to better deal with an ever-changing physical and social environment. And, artificial intelligence tools could mean you don’t have to fill in tedious paperwork, anymore. Ironically, given all the negative hype about automation and robots, AI has the potential to help you to feel more like a human being at work, and less like a cog in the machine.
Employee monitoring is a more controversial subject, however. UK businesses like JBrown have trialled software that predicts the ‘change-makers’ in work, analyses efficiency, and detects signs of overworking, with mixed success. The algorithms don’t give much weight to key parts of creative work, like thinking time. Employees could feel under pressure or not trusted in such an environment. What were we saying about being a cog in the machine again? A deeper thought must precede deployment of AI tools.
Plenty of social media automation tools now exist, giving you the chance to make your voice heard in a noisy space. ‘Doing’ social media as a business is time-consuming and can seem unrewarding. However, if you use AI tools to schedule, monitor and gather behavioural data on social platforms, you stand to gain vital information about your customers that can impact decisions you make. There’s even a chance that you’ve read a social media post in the past that was actually written by AI, and you couldn’t tell the difference!
Long gone are the days of keyword stuffing; that is, writing for the search engine rather than the human reader. Still, SEO specialists employ a bag of ever-changing tricks to try to optimise content for search engines. Google’s AI-powered Rankbrain, among other ‘signals’ that they use to determine content relevancy and quality, can understand the meaning of an article even if synonyms or natural language terms are used.
If you know about AI developments in SEO or you work with a specialist who does, your marketing team can update their strategy accordingly. You want people to find the information they need in the content you’ve produced.
With a bumpy history, natural language processing has finally developed to a stage where humans are relatively happy to talk to, and trust the information provided by, an AI-powered chatbot. 45 percent of millennials now expect a ‘good customer experience’ from a chatbot. Gartner predicts you will have more interactions with chatbots than with your spouse by 2020. Whether that’s good or bad news depends on your perspective.
All the top customer relationship management (CRM) software vendors are buying AI start-ups, focusing on internal development and leading the charge in using AI for customer service.
In particular, Salesforce has invested heavily in the industry. Their Einstein tool is a brilliant example to follow in using AI for business to benefit both their employees’ work and the customers’ experience. Among other things, it uses machine learning to predict business outcomes, such as customer lifetime value. It, along with another example found in Microsoft’s Virtual Assistant, also provides next step recommendations to optimise a customer’s journey. So, with ‘a little help from a friend’, your customer service providers can better sell products or services or resolve any issues.
‘AI will be an automated, vigilant sentry. And a reliable one, but only up to a point. A tireless assistant, but one occasionally in need of human supervision… Machine learning is a powerful pattern-matching technology. It's very good at identifying anomalies – variations from the norm, which is the tell-tale sign of most cyber-attacks.’ - Steve Lohr, Technology reporter at The New York Times
As attempts to exploit vulnerabilities in data security become more sophisticated, AI offers a way to protect data against future threats. If you’re promising customers that they can trust your business with their data, you’d better be good on your word.
As part of the ‘Carrefour 2022’ transformation plan, the multinational store chain is rolling out an artificial intelligence analytics solution, developed by SAS, to improve demand forecasting for their products. From their 18-month trial run, they’ve found some key benefits (beyond improved margins) that any business would count as a win:
- It made their teams more agile as they integrated new ways of working.
- Teams were better able to meet customer expectations due to efficiencies.
- As an iterative process, the supply chain became more efficient at every stage.
- They helped the environment by reducing waste and unnecessary transport of goods.
Similarly, Fujitsu use a method of machine learning called ‘dynamic ensemble learning’ for calculating highly accurate demand forecasts. Particularly, they have taken into account the drop in demand that often occurs after a new product is initially launched or a seasonal product appears on the shelves.
Over-stocking creates waste, which can often have a budgetary and environmental impact, and under-stocking can risk the credibility of a company. For example, in 2001, Nike experienced both. Poor forecasting meant they over-stocked on some shoes and failed to provide in-demand Air Jordans, which frustrated customers. They lost 100 million dollars in sales because of this error, which impacted their image at the time. Take them as the precedent NOT to follow.
Autonomous mobile robots (AMRs)
Finally, and perhaps controversially, we’ll talk about AMRs. Once, they were used by NASA in space missions, now they are in warehouses across our own planet. AI can be integrated with fleet management software to enhance navigation efficiency and reactivity via sensors and cameras.
Of course, with this comes the problem facing many workers in industries such as retail and manufacturing who ask: will automation take our jobs? It’s up to business decision-makers, and to an extent policy-makers, to innovate and ensure that humans can benefit from robots, AI and other disruptive technology, and not lose out because of them. How businesses react to these kinds of changes could determine if they are doing good, or just doing well.
Last, blue-sky thinking: profit AND people
These AI applications are being used or developed for imminent use today. 63 percent of businesses say that pressure to reduce costs will require them to use AI, with the market predicted to grow by over ten times from 2017 ($16.06 billion) to 2025 ($190 billion). Long-term, it’s clear there’s profit to be made…
… And, there’s good to be done. At Ditto, we believe in trustworthy and accountable artificial intelligence that provides expertise-led decision-making, which is the foundation for all good business. Contact us if you’re fellow do-gooder with an interest in responsible AI. We’d love to hear from you and help you on that journey.