I was honored to be recently invited to join a panel hosted by Aquent on how Artificial Intelligence (AI) is changing the marketing profession. And as I am often asked about this topic, in the spirit of sharing, here are some of my thoughts.
Chat-bots, virtual agents, robotics – oh my. I am often asked about which APIs and AI technology marketers should consider. But here’s the thing. Artificial Intelligence enables marketers to personalise and create more effective customer experiences, and improve ROI. At the core, artificial intelligence is all about technology which enables humans to make better, more informed decisions. But the first question is not which API do I need – but rather – what problem am I trying to solve.
Design thinking is a great place to start. Through empathy mapping to understand your customer or perhaps even your marketing team – how they think and feel and what they need and want – through to mapping of the problem statement and ideal end state – it is possible to clearly build a complete picture on what you are trying to solve and what outcome you are looking to derive. IBM in Australia and New Zealand has been piloting this in our own marketing department. In the first 3 months of operation we have already seen a 16% increase in CTR through improved advertisement placement.
With this in mind, AI can definitely assist in a number of key areas
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 the 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 – which you check next time to make better decisions – 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 devise, 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: 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: How to ensure simple and quick questions are supported, through repeatable models, whilst the valuable customer service team are leveraged for more complex questions.
How: There are a myriad of questions in which your customers ask, and will ask, about your business and solutions. Some of which can be answered quite easily and others which require the assistance of a human. Not all simple questions can be found on your web site – or – customers do not wish to scroll through pages of Q&A. 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 reserves the call centre wait time, for those who genuinely need to talk to one of your call centre customer service reps. UBank are doing this with their RoboChat offering, leveraging AI to help customers with their home loan questions. It is in market now. Check out their web site!!
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 – and it is easy to do!
How: Jason Grech is a gorgeous fashion designer in Melbourne. He spent 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 Ben Montague, the IBM Research Team, 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. Possibly weather impacts 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.
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 on 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. Imagine searching for an Avalanche Airbag system. Customers would search for ABS – which also stands for Abs of Steel, ABS Braking system and Australian Bureau of Statistics. Yes really! There are about 808 million results on google. However, if you were on the Northface web page as an example, customers would be looking for Avalanche Airbag system, a product offering in their range. However, by launching an Expert Personal Shopper built on AI, customers can type their requirements, using natural language – including ABS – 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 soooo much data on your brand. 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 – too preachy, too arrogant, too sarcastic. Building APIs to measure social sentiment analysis is possible using natural language and tone analysis to understand what content is relevant to your brand, what the sentiment is and helps you to make decisions on which posts to act on. And 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: Recruiting the best possible person in the team – to ensure you create a data driven culture of marketers curious and who embrace change
Solution: Recruitment solutions built on AI
How: Our wonderful partner Servian has created a compelling solution which, leverages tone analysis, social sentiment analysis – and more – to help your organisation profile the persona of the best possible fit for your organisation, and then find that person – leveraging their CV, social profiles, references and more. Tim Mannah, Head of Digital, at Servian shared this story with me “Imagine that you had to hire Sheldon from Big Bang theory – you are just not going to find that person on a CV. You need to look at so much more to find Sheldon – and AI makes that possible today.”
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. If means less time drowning in spreadsheets and reports and more time being creative!!!
This list will no doubt be outdated very soon. I would love to hear your examples of AI in marketing, and how you are embracing technology so you can love your job more!
♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦ ♦
Michelle Zamora is a Marketing Leader with a passion for business strategy, data, content, customer insights and engagement, and developing awesome talent. Views expressed are her personal, individual and unique perspectives, and not that of her employer.