Traditional marketing tools lack the flexibility, scalability, and comprehensiveness to address many of the challenges faced by modern companies. With growing digitization and an always-online audience, more marketing teams now require artificial intelligence (AI) to stay competitive.
Before rushing to hire a data science team, you need to evaluate the third-party AI solutions already available on the market. Many vendors use “AI” in their sales pitches, but lack credible research and engineering teams that can productize and operationalize cutting-edge AI research.
In this article, we feature companies with proven AI and ML expertise that are transforming marketing activities with state-of-the-art AI-driven solutions.
Appier is an Asia-based company established in 2012 by a group of true experts in artificial intelligence, data analysis, and marketing. The company offers several products that are driven by AI:
- AiDeal for tapping into hesitant customers with personalized deals in real-time. After leveraging this tool, Japan’s Urban Research Outlet increased conversion rate by 142% through smartphones and 116% through PCs, with revenue boosted by 70%.
- AIQUA for engaging customers based on their behavior inside and outside of your app or website. WIth this tool, Clovia, India’s leading lingerie and nightwear brand, increased the app and web checkout rates by 17% and 22% respectively.
- AIXON for unifying and enriching existing customer data as well as leveraging various AI models to predict customers’ future actions. After deploying this data science tool, CommonWealth Magazine increased subscriptions and purchases by 404%.
- CrossX Advertising for driving retention and ROI using better audience discovery with deep learning. By leveraging CrossX AI, Unilever’s AXE increased the click-through rate by 110% compared to industry benchmarks and expanded the original target audience by 50%.
Reactivating dormant mobile gamers with Appier (source)
The company’s team of AI and ML engineers is headed by professors from National Taiwan University and National Tsing Hua University with a proven history of cutting-edge research publications, including Hsuan-Tien Lin (Chief Data Science Consultant), Min Sun (Chief AI Scientist), and Shou-De Lin (Chief ML Scientist).
GumGum is an applied computer vision company. Their proprietary computer vision technology scans images, videos and the content that surrounds them across millions of web pages. With this information, GumGum’s customers can place contextually relevant ads where their users are most likely to see them. The company claims that their highly visible ad placements generate 7× more engagement and 37% higher brand lift than the industry averages.
For example, to increase the number of visitors to a leading hotel and resort chain, GumGum found travelers in the US planning their next getaway and served up highly visible, contextually relevant ads while these people were searching for content relevant to adventure and relaxation. As a result, the hotel chain got 148k extra visitors at a cost per acquisition of $4.73, which is 12.75× cheaper than other marketing channels typically offer.
Smart placement of ads by GumGum (source)
Furthermore, GumGum’s technology ensures that ads are published in brand-safe environments, avoiding web pages that talk about violence, illegal acts, disasters, etc. To identify pages that are suitable for advertisements, GumGum uses their NLP Threat Classification System and proprietary computer vision systems. Learn more about state-of-the-art approaches to brand safety from the talk by GumGum’s data scientist Sanja Stegerer.
3. IBM Watson
IBM is a tech company with a wide range of products and services. Applying AI to marketing and advertising is also one of its core areas of expertise.
In particular, IBM Watson Advertising offers a number of AI-driven solutions, including:
- IBM Predictive Audiences that leverages deep learning to score users based on the probability of their converting, making a purchase, viewing a video, etc. This analytics allows for better targeting precision and contextual relevance.
- IBM Watson Ads that enables intelligent and personalized conversations with consumers. Watson AI-powered conversational systems carefully listen to customer questions and deliver unique responses to each consumer. After deploying IBM Watson Ads, Best Western Hotels & Resorts achieved a 48% incremental lift in visits.
- IBM Advertising Accelerator that predicts the optimal combination of visual elements to drive the highest engagement and conversion for a given audience. It also uncovers target audience composition and preferences that can be applied to future campaigns.
Heuritech’s AI-powered platform helps leading brands around the world in predicting product trends. Their proprietary technology solution captures early signals from influencers and consumers by reviewing millions of social media images every day. The AI-driven product solutions include:
- Product intelligence for analyzing all the data shared by consumers and influencers and turning it into clear insights. For example, one of the clients asked Heuritech to find out which of the eight bags presented at a fashion show would be the most successful among consumers. After analyzing the number of shared images, the engagement rate, how often a photo was zoomed in on the product, and the number of brand mentions, Heuritech identified two bags that triggered the most enthusiasm. The brand used this data to get better sell-through with no overstock and no missed opportunities.
- Trend forecasting for anticipating and quantifying relevant trends and maximizing sell-through. For example, Heuritech helped one of its clients to spot black dad sneakers as a relevant trend.
In contrast to dozens of companies that only claim to use AI, Heuritech has true expertise in AI and machine learning. Its AI & Computer VIsion team includes 8 PhDs who help the company deploy state-of-the-art solutions at scale.
If you’re interested to learn more about computer vision assisting in the prediction of fashion trends, check out this RE•WORK talk by Charles Ollion, CTO & Co-Founder at Heuritech:
More talks from practitioners on business applications of AI are available in the RE•WORK AI library.
NetBase helps its clients to get in-depth customer experience analytics, social media analytics, and more. Their AI-driven product AI Studio is claimed to be the first fully-automated, next generation AI platform. It provides transparent and auditable results across all sources of customer experience data, including social media, reviews, and internal data sources.
AI Studio’s first innovation is Theme Discovery, a solution for linking semantically similar terms. It is pre-trained on millions of documents and automatically discovers themes that emerge from topic conversations. For example, you can rapidly discover themes emerging around your brand or industry, such as trending events, emerging crises, new promotions, reactions to media, rising influencers, and more.
iHeartMedia is leveraging NetBase’s AI solution for better curation, targeting, and amplifying content. This cooperation resulted in much more engaged users and a rapidly growing listener base.
The AI & ML team at NetBase is led by an experienced VP in Artificial Intelligence R&D, Karin Golde, who holds a Ph.D. in Linguistics from Ohio State University and has almost 20 years of experience developing NLP-based business solutions.
To enable AI-driven audience targeting and measurement, Quantcast has developed Q, purportedly the largest audience behavior platform for the open Internet, which directly quantifies over 100 million web and mobile destinations every day. It enables reflection of audience changes in real time and recognition of patterns that are not obvious to humans.
How does it work? The client shares with Quantcast the attributes of their ideal audience or identifies their existing audience by tagging. Quantcast builds a custom model using millions of data points available about these customers (e.g., demographics, pre-search behaviors, past purchases). Finally, it finds audiences that fit this profile and delivers the client’s message to them at the perfect time.
Discovering behavioral patterns with AI (source)
The effectiveness of Quantcast’s approach is supported by a number of successful case studies. For example, with their precise audience selection, Quantcast was able to bring prospects to IKEA’s website that were 16 times more likely to buy something than average IKEA site visitors. Next, MSC Cruises, the world’s largest privately owned cruise company, leveraged Quantcast’s solution to increase its paid and organic search traffic by 167% and get valuable insights about the potential audiences that the company had never considered before.
If you’re interested in learning more about how machine learning can transform marketing and advertising, check out this presentation by Quantcast’s Senior Modeling Scientist Sermetcan Baysal and his colleagues to see that for leading marketing teams AI is becoming an indispensable tool for audience discovery and targeting.
Salesforce provides a customer relationship management (CRM) platform that gives a single, shared view of every customer to all company departments, including marketing, sales, service, and others. However, it also offers a number of AI-driven tools aimed specifically at marketing and advertising. The Salesforce Einstein solution enables:
- a better understanding of the channels, messages, and content that resonates well with customers;
- giving every employee instant access to insights received with AI-powered analytics;
- better prediction of conversions and anticipation of marketing email performance;
- personalizing customer experience with individual product recommendations, messages, and even the timing of these messages;
- anticipation of customer behavior, including churn risks, delayed payments, lifetime value, etc.
According to the recent overview of AI in Salesforce Einstein, the tool applies machine learning for discovery and forecasting, natural language processing for sentiment analysis, intent discovery, and chatbot support, and computer vision for image classification and object detection.
Einstein’s prediction capabilities are successfully leveraged by Marriott hotels, Ireland’s leading recruitment agency Cpl, furniture seller Room & Board, and many others.
Dstillery positions itself as an applied data science company that builds the best audiences for its clients with custom AI models that are based on 10 million attributes. Their proprietary ProspectRank technology enables the scoring of hundreds of millions of candidate members inside and outside of audiences every day to identify the audiences that are the best match for individual brands.
With these carefully built audiences, companies can execute better-performing marketing and advertising campaigns. For example, Dstillery helped a leading footwear brand to build ad campaigns that brought $2.44 in revenue for every $1 spent. A streaming media brand leveraged the Custom AI Audiences by Dstillery to acquire high-value subscribers at a cost-per-acquisition 30% below competing offers.
In her talk at the 2019 O’Reilly Strata Data Conference, Melinda Han Williams, the Chief Data Scientist at Dstillery, explains how they apply deep learning to cluster customers into fine-grained behavioral segments, enabling smarter marketing decisions.
Affectiva leverages computer vision and speech analytics to recognize human emotions. Their technology has many applications, and market research is one of them.
With Affectiva Media Analytics, you can optimize your brand content and media spend by measuring unfiltered user responses to ads, videos, and TV shows. Quality emotion data on customers’ response is critical for predicting key advertising success metrics, including sales growth, purchase intent, brand recall, and likelihood of sharing.
To provide you with this data, the Affectiva Emotion AI technology measures viewers’ moment-by-moment facial expressions of emotion while they watch your video. The results are displayed in an easy-to-use dashboard. See the demonstration video below.
When helping the Mars brand to optimize their advertising strategy, Affectiva developed a large research study to tie facial reactions and emotional responses to sales effectiveness. They captured the facial reactions of over 1,500 participants from France, Germany, the UK and the US while they were viewing 200+ ads. After analyzing the emotional responses of all participants and combining this data with a self-report survey, Affectiva was able to predict short-term sales with an accuracy of 75%.
The Affectiva team is led by Rana el Kaliouby, a recognized tech leader with a research background from MIT Media Lab. See her recent talk on pioneering artificial emotional intelligence.
Cognitiv applies its deep learning-based technology NeuralMind for advertising optimization.
It’s always hard to predict the performance of an ad, especially considering that the same advertisement can work for one set of customers but not for another. At the same time, deep learning is perfectly suited to help marketers take advantage of huge datasets and make smart data-driven decisions. In contrast to basic machine learning techniques, such as logistic regression or decision trees, neural networks can fully leverage the data and recognize complex patterns, which means deep learning can help to discover revenue opportunities missed by other analytical tools.
The solutions offered by Cognitiv include:
- Cognitiv Performance to improve the ROI of bid-based campaigns with autonomous deep learning buying algorithms.
- Cognitiv Incremental to improve the measurement of ad effectiveness by predicting whether an impression will deliver conversion and whether this conversion is due to additional marketing.
- Cognitiv Audience to identify high-value users more effectively and efficiently.
- Cognitiv Search to optimize Google Ad campaigns with automated bidding and adjusted creatives.
The Cognitiv team is headed by deep learning experts, such as Marc Hudacsko, the company’s CTO, and Aaron Andalman, its Chief Science Officer. Aaron holds a Ph.D. in Neuroscience from MIT and works as a postdoctoral scholar at Stanford, researching neural networks and data analysis techniques.
This article was originally published on TOPBOTS.