Understanding AI: the need-to-know terms for pharmaceutical marketers
From retail to finance to aviation, the applications of artificial intelligence (AI) are many and diverse, with 80% of large companies adopting some form of machine learning and AI.
Its dominance in tech news extends to the pharmaceutical sector, which is by no means immune to the charms of AI, with Novartis, Pfizer, GlaxoSmithKline and AstraZeneca just some of the firms signing deals and forming collaborations in the area.
To date, much of this publicly announced work has been focused on drug development in therapy areas such as diabetes, multiple sclerosis, kidney disease and many others, but the possibilities are endless – especially within pharmaceutical marketing.
Successful marketers will be the ones who know how, and why, to apply AI and machine learning to their products, teams, plans, strategies and tactics to keep pace with the changing marketing world.
As the technology is increasingly embedded in our everyday lives, it’s clear that, rather than being a flash in the pan, AI offers increasingly sophisticated ways for marketers to engage with the data revolution.
This is particularly true with respect to natural language processing and its applications in areas such as social media monitoring, chatbots, customer service automation and content marketing strategy, to name but a few.
As the technology continues to mature, the possibilities for AI and machine learning in pharma marketing seem endless, but to successfully implement solutions means getting to grips with the growing cavalcade of jargon associated with them.
So, to help pharmaceutical marketers navigate a path through terms such as ‘ambient intelligence’ and ‘unstructured data’, InTouch has put together a glossary of 15 key AI terms to help bring you into the conversation and engage fully with cutting-edge tech activities.
The AI glossary below is part of Intouch’s spotlight on AI in pharma and provides definitions of the key concepts driving the modern tech revolution.
A computer program governed by a specific set of rules that allows it to perform complex, labor-intensive tasks like calculations, data processing and automated reasoning … so we human marketers can focus on strategy and creative.
Ambient Intelligence (am•bee•uh•nt | in•tel•i•juh•ns)
“Smart” devices like Alexa, Google Home and Apple HomePod are sensitive and responsive to the presence of people. They hang out in the background and wait for you to ask for their help.
Artifical Intelligence (ahr•tuh•fish•uh•l | in•tel•i•juh•ns)
A computer system that can gather data and make decisions and/or solve complex problems. Intouch’s proprietary artificial intelligence (AI) engine built specifically for pharma, Cogntive Core, is one such example.
Augmented Reality (awg•ment•ed | ree•al•i•tee)
Yep, like Pokemon Go, or the Intouch app, In My Eyes, AR helps you see — using your phone or special goggles — computer-generated things that aren’t there in the real world.
Behavioral Informatics (bih•heyv•yer•ol | in•fer•mat•iks)
The use of technology/devices to detect and measure human behaviour to gain insights. For example, searching Google for “large dog breeds” tells data-collection folks that you may be thinking of adopting a dog, which means there’s a good chance you’ll start seeing online ads for pet supplies. Our programmatic media team uses these kinds of insights to develop more effective targeting protocols.
Big Data (big | dey•tuh)
This is the massive amount of information we now generate about ourselves – our interests and habits – as we move through the digital universe. Some say the term ‘big data’ should be retired, because so much data is collected these days that all data is now part of big data.
Chatbots are programs – like Apple’s Siri – that simulate human conversation, using response workflows or artificial intelligence to interact with people based on verbal and written cues. Chatbots can be the frontline of communication between brands and their users. Intouch’s Cognitive-Core-powered chatbot, Ruby, can help users get information on medications and treatment plans, help patients set doctor appointment reminders and more.
A group of people that share common characteristics such as age, parental or marital status, hobbies and pretty much anything else you can think of. AI programs can identify clusters and reveal patterns that help marketers target groups of people with common characteristics.
Deep Learning (deep | lur•ning)
A more advanced branch of machine learning, where a computer teaches itself with only minimal amounts of programming. With deep learning, marketers can make predictions about consumer behaviour.
Image Recognition (im•ij | rek•uh•g•nish•uh•n)
AI looks for patterns in images, and the tech is scarily good. As of 2016, the error rate was less than 3%. A January 2017 article in Nature described an AI system that could diagnose and classify skin cancers just as well as board-certified dermatologists.
Machine Learning (muh•sheen | lur•ning)
Machine learning teaches a computer to find functions – equations that work not only for the examples that it has, but for unknown ones in the future. Machine learning teaches a computer how to predict.
Natural Language Processing (nach•er•uh•l | lang•gwij | pros•es•ee•ng)
Natural language processing is a way for computers to analyse, understand and derive meaning from human language. Where can NLP be used?
- Adverse event detection
- Sentiment analysis
- Text analysis
- Text generation
- Text summarisation
Neural Network (noo•r•uh•l | net•wurk)
This is essentially a two- (or more) heads-are-better-than-one approach to problem solving. Neural networks – designed to be similar to the human nervous system and brain – help AI solve complex problems by splitting the work into levels of data. These networks can be used to recognise handwriting or faces, for example.
Turing Test (too•r•ing | test)
Developed by computer scientist Alan Turing in 1950, this was a test to determine whether a computer could think. If a human interacting with it believed they were talking to another person – not a computer – the test was considered a success.
Unstructured Data (uhn•struhk•cherd | dey•tuh)
This is what it sounds like. Disorganised chunks of data that appear random and unconnected. Examples of unstructured data include email messages, social media posts, photos, audio files, text messages, satellite images and webpages.
- Download the AI glossary for pharma marketers