We hear that Google is using AI in search, Facebook uses it for facial recognition and Netflix is using AI to ‘conquer the world‘.

These examples are all very interesting, but they do leave many wondering what exactly AI is and how can they apply it, now, to their everyday marketing tasks?

To help marketers understand AI and how it applies to our craft, Econsultancy recently held a Digital Outlook event in Singapore and invited marketing AI expert Deborah Kay, Founder of Digital Discovery, to give an overview of the state of the art.

Econsultancy’s Marketer’s Guide to Machine Learning and AI

Helpfully, Ms. Kay provided a summary of the four most exciting areas of AI for marketing as well as many examples of how AI is being used in the real world. Below is a summary of her presentation.

Define ‘AI’

To start things off, Ms. Kay wanted to make it clear what people mean when they talk about AI.

“Artificial intelligence (AI) is the ability for computers to perform tasks that, in the past, only humans could do.”

That is, AI is a technology which allows computers to conduct analysis, write copy and even serve customers. Additionally, unlike humans, computers can draw upon vast data resources which would be difficult, if not impossible, for people to explore while conducting these tasks.

So how exactly can AI be applied to marketing? Ms. Kay told attendees that there are four main ways in which AI is being developed to handle tasks traditionally carried out by marketers.

outlook event

1) Computer vision

While everyone is well-aware that computers have been able to take and store photos for decades, with AI, computers can start to make sense of the people and objects in the images.

So how does this help marketers? One example offered by Ms. Kay is that AI can be used to tag and classify images, making them more available to site searchers and web crawlers for SEO purposes. As a marketing consultant for the SG Sailing Club, Ms. Kay used AI to tag photos from regatta observers and allow participants to find photos of themselves without any manual work.

Computer vision AI can also be used, as platform deep.social did, to categorize influencers according to their demographics, brand affinities, and interests.

2) Natural language generation (NLG)

Another exciting application of AI which is relevant to marketers is the steadily-improving ability of computers to convert data into conversational-style language.

One current example of this, available to all digital marketers, is Google Analytics (GA). Located in the ‘Analytics intelligence’ toolbar, GA reports, in English, noteworthy events. Insights include topics like ‘The bounce rate decreased on some landing pages’ and ‘More users returned to your site in December’.

The company Narrative Science also provides a service, Quill Engage, which takes GA data as an input and produces natural language as an output. Quill Engage, however, offers information which marketers can easily cut-and-paste into a weekly management update, saving hours of time and effort.

quill engage

A Quill Engage sample report

Another company using NLG to help marketers is Automated Insights whose Wordsmith product helps generate engaging, unique copy for product descriptions and categories at scale.

Finally, NLG can also be used to increase a marketer’s creative output. Phrasee, helmed by Econsultancy contributor Parry Malm, helps marketers craft attention-getting emails and social ad copy which maintain brand voice and often sound ‘more human’ than those actually written by humans, themselves.

3) Conversational chatbots

While AI-based assistance, or chatbots, may have fallen out of the spotlight lately, marketers were urged by Ms. Kay to take another look at their potential.

In Singapore, over 100,000 commuters are now plugged into the ‘Bus Uncle‘ platform which offers bus arrival times via Facebook Messenger. Bus Uncle accepts natural language questions (‘when will the 198 get here?’), asks English follow-up questions (‘where are you?’) and responds using amusingly colloquial ‘Singlish’ (‘7 mins go make soft boil egg’).

Ms. Kay also provided a case study of a chatbot she was instrumental in developing for the Singapore Sailing Federation, Sammy, which offered sailing fans several services including:

  • Responding to queries about a regatta
  • Providing up-to-date micro-climate weather reports
  • Acting as a central repository for photos of a sailing event

Each of these applications made life easier for the organisation’s marketers, improved the experience for event attendees, and resulted in 3x the number of feedback survey respondents than previous regattas, with 97% saying that they wish to keep the chatbot for future events..

4) Personality insights

A relatively new AI application, also useful for marketers, is the analysis of publicly-available social media and blog posts, resulting in personality profiles of the writers.

Frrole.ai has commercialized this technology to provide brands with audience intelligence and make suggestions to customer service teams about how best to respond to people based on the language they use on social media.

IBM Watson Developer Cloud, too, offers AI-generated personality insights through analysing Tweets – which marketers can try out themselves at the IBM Watson website.

jeff-personality-portrait

Leveraging this technology can help marketers both identify characteristics of their customer base as well as improve customer service.

5) Combinations

Finally, Ms. Kay offered attendees a few examples of how AI technologies could be combined to offer powerful services which would be difficult for humans to emulate.

Chatbot + Personality insights

A chatbot which identifies clients who are using atypical ‘unhappy’ language and redirect conversations to customer service personnel.

Chatbot + computer vision

A chatbot which can take photos as chat instead of making people type in details, such as a product SKU or a loyalty card ID.

Personality insights + natural language + chatbot

A chatbot which can change the language it uses depending on the personality of the person it is chatting too.

With the technology innovations now available, the possibilities of harnessing AI for marketing, she concluded, are endless. It is up to marketers now to explore the AI universe and find which solutions are best suited for their brand.

outlook-presentation-sg