10 examples of AI-powered marketing software

Some of the solutions are fairly novel, and others have been around for a while.

Obviously, this isn’t an exhaustive list, but it gives an idea of the maturity of machine learning.

1. Email copy – Phrasee

Appropriately enough for a language optimisation product, Phrasee’s website describes the product with admirable clarity:

“Phrasee is language optimisation software. It gives you human-sounding, machine-optimised marketing language that gets you more opens, clicks and conversions.

“Better results come from better use of language combined with better statistical analysis. Until Phrasee, this was impossible for most marketers.

“Phrasee’s artificial intelligence algorithms generate your subject lines, body copy, and calls-to-action. It looks at hundreds of emotions, sentiments and phrases and predicts what your audience will respond to.”

“The more you use Phrasee, the better your results get.”

Econsultancy blogger and Phrasee founder Parry Malm explains that “in the near term (say 10-20 years) machines won’t be able to outperform humans when it comes to creating long-form text. What Phrasee can do is outperform humans with small, structured language sets such as subject lines.”

The Phrasee team received £1m seed investment in July 2016, and is currently the only SaaS language optimisation product on the market.

Incidentally, Parry has written some great articles on email marketing, which you should read immediately.

2. Conversion optimisation – Sentient Ascend

We covered Sentient Ascend on the Econsultancy blog earlier this month (Could AI kill off the conversion optimisation consultant?).

Machine-learning algorithms allow for more efficient multivariate testing. Scores of website features can be tested, requiring less traffic than traditional testing.

Sentient’s website explains: “Our patented AI solution mimics biological evolution, enabling it to quickly learn, adapt and react to determine the best performing design from the building blocks you provide.”

There are a number of brands onboard, with underwear brand Cosabella generating 35% more conversions than the control when testing 15 different changes to the homepage header, category page, product page and cart layout.

The Sentient Ascend interface

3. Virtual agent – Watson

Watson does a whole bunch of stuff, of course. From a virtual agent to search, analytics and unstructured text analysis.

Let’s look at the virtual agent. According to IBM “it offers a cognitive, conversational self-service experience that can provide answers and take action.”

The agent is pre-trained with industry and domain content, but is customised to fit the user’s needs, content and brand. More from Watson.

4. Purchase Recommendations – Grey Jean’s Genie

Grey Jean claims that Genie can predict a customer’s next likely purchase with “up to 72% accuracy in a category”.

The personalisation platform uses a whole host of user data, from online and offline purchases and loyalty programs, to CRM, social media and website behaviour.

Demographics and income level are also used in modelling the best offers to present to customers, in the most efficient channel and at the best time.

Deals can be delivered via channels such as web, geo-targeted push notification, social ads and email, and the tool can be used for personalisation (in the case of a recognised user) or for behavioural segmentation.

It’s worth noting that CEO Craig Alberino told Martech Today that the platform is not specifically set up to match offline and online selves. Offline data is primarily used to assist geotargeting and the building of behavioral profiles.

5. Lead generation – DemandBase’s DemandGraph

DemandGraph uses Demandbase’s business records alongside publicly available information such as newswires, regulatory filings and social media to profile potential clients.

Analysing this unstructured text as well as billions of web interactions from B2B buyers gives a good estimation of what prospects are looking for, at what time, and through which decision maker.

Users are given an overview of potential clients including key information such as corporate structure, decision makers and relevant content they have published.

The software can generate custom messages tailored to this information, which a user can take as the basis for their own contact.

DemandGraph’s website claims: “Mapping the relationships between companies is far superior as a predictor of a future relationship between two companies.

“DemandGraph provides companies with a trusted and accurate repository of information which they can use to guide conversations, better predict future business behavior—and with accurate precision—identify and target their next customer, supplier or partner.”

6. Performance marketing – DataXu Mobile Optimizer

Programmatic advertising makes use of machine learning in its targeting of users most likely to click on a given advert.

I thought I’d include a technology that focuses on mobile specifically. DataXu’s Mobile Optimizer uses the company’s cross-device tracking technology and machine learning to “accelerate the growth of app installs and increase engagement of users across all their devices and apps.”

Device matching itself (the mapping of mobile device IDs to cookie data) is done with machine learning, and so is the optimisation of targeting (based on CRM, purchase history and behavioural data).

Ed Montes, Chief Revenue Officer for DataXu, said in a release that: “The most innovative marketers are recognizing the value of mobile applications and the data associated with them as the strongest bond between brands and their customers from both a customer experience and marketing intelligence standpoint.”

7. Cloud Machine Learning Platform - Google

Okay, this isn’t marketing software per se, but Google’s cloud platform deserves inclusion here, not least because Alphabet has just announced the creation of an AI unit for Google Cloud led by Stanford University intelligence professor Fei-Fei Li.

Google Cloud ML Platform provides “machine learning services, with pre-trained models and a service to generate your own tailored models.”

This is the same platform that Google uses for Photos (incredible image recognition), voice search, Translate and Gmail’s Smart Reply. Customers can now bring its power to their own business applications.

8. Salesforce Einstein – sales, marketing, service, community, commerce, analytics, and Internet of Things

Salesforce rolled out its Einstein AI in September 2016 and the range of applications is impressive.

There are various flavours of product available with AI enabled in each Salesforce Cloud (listed in the subheader above). Some of these new features will come with no added charge, whereas others will be paid for.

Examples include:

Sales Cloud: Predict which leads are more likely to become sales, alert users to contacts that might be considering a rival service.
Marketing Cloud: Analyze images on social media, perform customer sentiment analysis, and make suggestions about targeted marketing.
Service Cloud: Case classification.
Analytics Cloud: Predictive analytics, recommendations and automated rules optimisation.

As it sounds, Salesforce is going big on AI, and has also formed a Salesforce Research group.

This group will be a 175-strong team of data scientists looking at deep learning, natural language processing, computer vision and more.

Copy from the Salesforce Einstein web pages

9. Messaging – Boomtrain

Boomtrain Messenger offers multichannel recommendations but with a bit more sophistication.

An online chat tool can be implemented using past user behavior to “engage customers in intelligent conversations at just the right time.”

As you can see from the screenshots below, it looks fairly smart, and offers an alternative to yet another email (particular as it can engage in real time).

10. Account-based marketing – YesPath ABM

YesPath ABM automates segmentation of website content in the targeting of key accounts (and the key people within those accounts).

JavaScript in the site header allows the tagging of content according to topics. The platform then presents questions and content to the user according to a continuously optimised user profile.

Upon its release in March 2016, YesPath compared its AI approach to Facebook’s, where an algorithm determines the best content for your news feed.