July 31, 2017
Categorized: Artificial Intelligence
Share this post:
Written by: Michelle Zamora, Head of Marketing, Banking, Insurance, Cognitive, Analytics, Software, IBM Australia
Chat-bots, virtual agents, robotics – oh my! I am often asked about which APIs and AI technology that can improve customer experience. Artificial Intelligence enables businesses to personalise and create more effective customer experiences and improve ROI. At the core, artificial intelligence is about technology which enables humans to make better, more informed decisions. The first question is not which API do I need – but rather – what problem am I trying to solve.
With this in mind, here are some key areas AI can definitely assist in.
Challenge: My customers are waiting on hold to speak to the customer service team for long periods of time. They are regularly hanging up without getting the advice they need.
Solution: Ensure simple and quick questions are answerable with, through repeatable models, whilst the valuable customer service teams are freed up to spend more time on complex questions.
How: There are a myriad of questions in which your customers can ask, and will ask, about your business and solutions. Some of these can be answered quite easily and others require the assistance of a human. Not all simple questions can be found on your web site – or – customers might not wish to scroll through pages of FAQs. A virtual agent, built on AI, will understand natural language in context to your business. A customer can ask questions and receive responses, often fast tracking to purchase from there. This reduces the call centre wait time for those who genuinely need to talk to one of your customer service reps. UBank are doing this with their RoboChat offering, leveraging AI to help customers with their home loan questions.
Challenge: The ability for your customers, to be able to navigate and stay on your web page, for all their requirements
Solution: Natural Language interpretation
How: Your customer jumps on your web page and with the search bar, searches for a product that is relevant to your portfolio of offerings. However, when your customer searches for the product it takes them to every other product but yours. The customer has to invest effort and time to find the product they are looking for – often outside your web page. Meaning they are exposed to more competitive products. It is too hard to buy – so they don’t. If your site search capabilities are inadequate, it could impact your sales – what if AI could help? If you were on the Northface web page for example, finding the right product is a few simple questions away. By launching an Expert Personal Shopper built on AI, customers can type their requirements, using natural language and they can be certain that they will receive answers to their question – in context – every time they ask. When 70% of shopping carts are abandoned online today, this is a competitive advantage.
Challenge: The ability to understand the social sentiment of your brand
Solution: Social sentiment analysis
How: There is So much data on your brand, but how do you tap into all of this to truly understand the tone of the sentiment? What if you want to assess your content and how it could be perceived – before it even hits the market? What if you could find out if it was too preachy, too arrogant or sarcastic? Building APIs to measure social sentiment analysis is possible by using natural language and tone analysis to understand what content is relevant to your brand and what the sentiment is, whilst helping you to make decisions on which posts to act on. It allows you to also review your content before you even hit the post button (I do this before I post my blogs – it definitely streamlines the editing process!)
Challenge: My digitally savvy customers do not know all the ways in which we can help support their business.
Solution: Improve digital content response rates with limited marketing funds
How: The implementation of artificial intelligence, to assist the marketing team select the best keywords, the best media to fund, and even the best time to place the ads.
Working with digital agencies to bid for ads, select the most appropriate media based on historic data and hoping for the best possible outcome (conversion to clicks, page views, responses, revenue generation) is now become an antiquated process. Programmatic Bid Optimisation is here and is leveraging AI to optimise digital media. It is improving who to target, the time to target them, and ensuring you are bidding the right ad at the right time. Imagine knowing – not estimating – which size ad works better and the best time of day to place that ad. Imagine being able to modify throughout the day, varying ad sizes and time, to maximise CTR. Now add in factors such as frequency, browser, location device, language. Throw in elements like weather (which can impact everything from what we buy, to when we buy, to where we buy). All these factors can make it more difficult to decide where, what and when to place an advertisement. With AI we can move from bidding to knowing – and ultimately improving the ROI and certainty of impact of digital marketing.
Challenge: How can I help my creative team spend less time reviewing and assessing data, so they can spend more time delivering creative responses and increase revenue
Solution: APIs built on AI can be easily customised to your organisation at a low cost.
How: Jason Grech is a fashion designer based in Melbourne. He would spend a 12-week lead-up to a fashion show looking at fabrics in stores, looking at Vogue fashion guides, attending fashion shows and trying to determine what styles, colors and fabrics he would use. This would take up the majority of his lead time, leaving him minimal time to design and create. When Benjamin Montague, Ogilvy and I spent time with Jason, we learnt he was inspired by architecture and European trends. He loved black and white. He loves Melbourne and adores ensuring his clothes flatter the bodies of the women who wear his gowns. We created two tools built on AI (APIS) which utilised IBM Watson. The first, The Visual Discovery Tool using Visual Recognition APIs, looked at photos of the architecture that inspired him and mapped this to silhouettes from historic fashion shows, and, using detailed image understanding, helped to recommend dress styles. The second, the Zeitgeist Tool, again using Visual Recognition APIs, analysed trends of colours and styles from Instagram images, to help predict colors, styles, necklines, cuts and fabrics. In the end 12 beautiful dresses were created in record time, going from inspiration to storyboard in just 4 days, allowing the majority of time to be spent creating. The color of trend – Lilac. This was way out of Jason’s comfort zone – but he trusted the guidance, trusted the process and combined with his creative genius, his best ever collection hit the market in September last year! And recently – Jason made his way into Vogue Italy – Amazing!
Challenge: The marketing department are creating leads which are not converted to revenue for the business
Solution: Improve the lead management process, ensuring the right leads are passed to the right person or business partner at the right time
How: AI can leverage data to better predict which leads will close, scoring leads to prioritise for your sales team. Further, it can in real time, assess the performance and capability of sellers – including external business partners – to ensure that you are placing your best leads with your best and most appropriate seller, for maximum chance of closure. Imagine knowing which route to market is more effective, for which leads, at which time of day. Even possible impacts on the lead quality and likelihood to close, or the performance and time it takes to close. AI can do this today to improve lead management.
The examples grow and expand every day. These are just a few of the ones I prefer to start with. AI for marketers is quite simply, fun. It means less time drowning in spreadsheets and reports and more time being creative. This list will no doubt be outdated very soon, as the capabilities of AI continue to grow and more exciting tools and API’s emerge.
Get to know IBM Watson better Here.