Eight by Eight

eCommerce & Artificial Intelligence Consulting

Start Increasing Sales Now!

Get Our Free Newsletter

 

    You are here: Home / Glossary
    • Alexa Built-Ins are third-party devices that allow you to access the Alexa voice service.
    • An Alexa Flash Briefing is an Alexa skill that comes with your Amazon smart speaker. You can customize your Flash Briefings to get personalized news stories, weather reports, and updates from your favorite sites. Alexa will read or play your Flash Briefings in the priority order that you set.(...)
    • Shortcuts designed to save you time by grouping together actions. (This saves you from asking for each one individually.) You can activate your routines with your smart devices, via voice, time of day, and more.
    • Alexa Skills are voice-driven apps for the Alexa. Users use Skills to check the news, listen to music/podcasts, play games, order forms, control cloud-connected devices, and more. There are tens of thousands of skills in the Skill Store. Skills can be used on Alexa-enabled devices (Amazon(...)
    • Code words and phrases that folks use to keep their videos, posts, and comments taken down by algorithmic content moderation systems. You see it a lot on TikTok, Twitch, YouTube, and Instagram. (Examples: Panini = pandemic. Unalive = dead. Seggs = Sex. Corn emoji = porn.)
    • Amazon Personalize is a fully-managed machine learning system that helps you deliver customized recommendations to your customers.
    • AMP for Email allows senders to include AMP components inside rich engaging emails, making modern app functionality available within email. You can read more about them here.
    • The attribution of human characteristics, emotions and/or behavior to an object, animal, plant, or supernatural being. The old school example is Winnie the Pooh. These days, it’s a hot topic in robotics.
    • Artificial General Intelligence (AGI), also known as Strong Intelligence or Deep Intelligence, is a (currently) hypothetical machine that can perform ALL the tasks that a human can. Intended to think on their own, AGI systems are built with the human brain as the archetype. Like humans, AGI(...)
    • Artificial Narrow Intelligence (ANI), also known as Weak Intelligence, is a collection of technologies that rely on algorithms and programmatic responses to simulate intelligence, with a focus on a specific (narrow) task.  ANI is limited in scope and very goal-oriented. Think recommendation(...)
    • Artificial Super Intelligence (ASI) is the hypothetical AI that surpasses human intelligence. The machines become aware and more capable than humans on every level. They’ll think and act like humans (but better) and they’ll have emotions, feelings, beliefs, needs, etc.  Plus, they’ll have(...)
    • Astroturfing is the deceptive practice of presenting comments, reviews, messages, and other information as if they came from the public instead of an organization.
    • Allows us to superimpose a computer-generated image on a picture of ourselves and/or our surroundings to create an enhanced (composite) view.
    • The computer’s ability to identify and process human voice. Automatic Speech Recognition (ASR) allows us (humans) to interface with computers by converting speech to text, in real time.
    • The length of time a user/visitor stays actively on your site. Average Active User Session is not the same as Average User Session. Most analytical packages only show the latter. Average Active User Session is calculated by dividing the user session by the number of pages to get the average(...)
    • Back propagation is an algorithm that is used to calculate derivatives quickly. It helps improve the accuracy of predictions in machine learning.
    • A statistical inference used to update the probability of a hypothesis as more information/evidence becomes available. “Bayesian” comes from the frequent use of Bayes’ theorem.
    • In Greek legend, the Centaur is half-man half-horse. In the ever-changing world of AI, the Centaur is half-man, half-machine. Well, technically the human brain combined with computer intelligence.
    • Churn staging (aka churn prediction) is the method of determining when a customer will stop doing something with you. It’s often used for determining your customer/buyer attrition rate – in other words when a customer will stop doing business with you. However, it can also be used to measure(...)
    • Computer Vision allows computers to "see." Using Artificial Intelligence and Machine Learning, photos, videos, and other images are converted into meaningful data that The Machine can understand/interpret.  
    • Optional feature that enables your Google Home Speaker to continue listening for another command for a few seconds or until you say "thank you."
    • Conversational Repair is the process speakers (humans & computers) use to detect and resolve issues/errors around speaking, hearing, and understanding.
    • The change in the quantity demanded for one product due to the percentage change in the price of another product. Often used in determining whether to increase or decrease your prices. Also helpful for figuring out viable substitutions.
    • From OpenAI, DALL-E 2 is an artificial intelligence program that creates realistic images and art from text descriptions. The best way to learn about DALL-E 2 (and Imagen) is to experience them.
    • Data decay refers to the gradual deterioration/corruption of data quality in sales and marketing databases. Also known as Data Rot.
    • Data exhaust is the trail of data left from anything you do online (buy, browse, watch, write, etc.) It’s all the cookies, temporary internet files, and so on. It’s often referred to in discussions about predictive analytics. Or frankly, privacy.
    • Deep Learning is a subset of Machine Learning which is a subset of Artificial Intelligence. Deep Learning works with multi-layered neural networks. It uses both structured and unstructured data for training.
    • Deep Learning + Fakes = Deepfakes, a type of artificial intelligence used to create convincing auto, video, and image hoaxes. The quality of deepfakes has been getting better and better and it’s becoming harder (although not yet impossible) for people to spot.
    • Tells you WHAT already happened.
    • Tells you WHY something(s) happened.  
    • Digital Assistants are advanced computer programs that understand and respond to voice commands. Digital Assistants are available on many smartphones, computers, tablets, and standalone devices. They use special microphones and software to listen for specific spoken commands from users to(...)
    • Techniques that minimize the number of input variables in your dataset(s).
    • Driver Analysis is a subset of Predictive Analytics that tells you what makes something happen. It helps you understand what combination of efforts/variables lead to a particular goal/outcome. Also known as Importance Analysis.
    • Duplex AI is a feature of Google Assistant that uses AI to carry out “real-world tasks” over the phone. Example: calling a business on your behalf and scheduling an appointment or making a restaurant reservation. It uses a realistic-sounding artificial voice.
    • Dwell time (aka Primary Evaluation Page) is the length of time a user spends looking at your page after they’ve clicked on a search results page in Google/Bing/etc. but before clicking back to the SERP results. It indicates to the search engines whether or not your content is useful.
    • An umbrella term used to describe augmented, virtual and mixed realities.
    • Few Shot Learning (FSL) is a type of Machine Learning where the Training Set has limited examples. (The Machine is expected to learn something from just a few examples.) Few Shot Learning is used in Computer Vision, Robotics, Natural Language Processing, Acoustic Signal Processing, and much(...)
    • Friction is any sort of resistance that prevents a user from moving to the next step in their journey. Friction has an increasingly large impact on how search engines find/traffic your site. It’s, by far, one of the biggest challenges marketers face today as it impacts the quality of(...)
    • Google Actions extend the functionality of Google Assistant. Equivalent to Alexa Skills, they’re like smartphone apps for your voice assistant.
    • Designed by OpenAI, GPT-3 is a language prediction model that uses deep learning to create human-like text. (In a nustshell, GPT-3 writes text that sounds like a human wrote it, kinda sorta sometimes maybe.)
    • Developed by Washington Post, Heliograf was designed to “enhance story-telling for large-scale, data-driven coverage of major news events." Also known as the first robot reporter.
    • Hyperpersonalization uses artificial intelligence and real-time data to create highly targeted, uber-personalized content for every individual/visitor on your site. It’s old-school personalization but on steroids.  
    • Using deep learning and computer recognition, image caption generation is a process of recognizing the context of an image and automatically labeling/annotating it with appropriate captions.
    • Developed by Google’s Research Brain team, Imagen creates art based on written text. Google describes it as a “text-to-image diffusion model with an unprecedented degree of photorealism and a deep level of language understanding.”  The best way to learn about things like Imagen and DALL-E-2 is(...)
    • ImageNet is a large visual database of millions of images designed for computer vision research.
    • Programs that can make decisions or perform actions autonomously to achieve goals.
    • Intent is what a user actually means when they say or write something. It’s different from an Utterance which is what a user says.
    • An invocation name is the word or phrase that users say to trigger (invoke) a custom skill.
    • Long-tail keywords are highly-targeted search phrases that are more specific to the user's intent. They typically have longer word counts and get less search traffic but have higher conversions. Long-tail keywords have been around forever. We’re talking about them more now because of Voice(...)
    • A setting that determines how far back in time to target visitors for a campaign.
    • Programming that assumes the prejudices and biases of its data or its creator.
    • Machine Learning is a branch of Artificial Intelligence that enables software to become better at predicting outcomes without being explicitly programmed.  
    • Mixed Reality (MR) is essentially an upgrade to Augmented Reality (AR.) AR enhances a user’s perception of the real world. Mixed Reality blends the real world and an imaginary world together allowing users to interact with the combined worlds (physical and virtual.)
    • Modality refers to the way something is experienced. Multimodality refers to the combination of multiple types/models to create meaning. For example, a voice interface + a screen.
    • Natural Language Generation (NLG) enables computers to produce natural written or spoken language.
    • Natural Language Processing (NLP) is what happens when computers read (human) language. It analyzes and processes large amounts of natural (human) language data into text that a computer can understand.
    • Natural Language Understanding (NLU) focuses on machine (computer) reading comprehension.
    • A neural network is set of algorithms modeled on the human brain and nervous system. They allow The Machine to recognize patterns and solve common problems in the fields of AI and machine learning.
    • Next Best Offer (NBO) allows you to pinpoint which offer is best for an individual. Artificial Intelligence has made determining NBO more robust than ever -- not only will it identify which offer(s) a visitor is most likely to respond to, it can pinpoint when and how you should present the(...)
    • Next Best Product (NBP) identifies which product(s) a user is most likely to buy next. It's been used for many years to increase revenues, improve buyer frequency, and increase lifetime value. Artificial Intelligence has made calculating NBP easier than ever, figuring out what each customer(...)
    • Next Best Product (NBP) refers to identifying and personalizing the product(s) best suited to a particular individual to get them to convert. (Or inquire, adopt, sign-up, whatever it is that you want them to do.)
    • An outro is the closing section of an ad, a piece of music, a TV or radio program, etc.
    • An Owned Virtual Assistant is a Virtual Assistant built out in a brand’s unique voice.
    • PaLM is Google AI’s 540-billion parameter, densely-activated Transformer language model.
    • The Pareto Principle (aka the 80/20 rule) says that for many outcomes, 80% of the consequences come from 20% of the causes.
    • A series of events experienced by a visitor during a session on your site. Paths include pages they visit, actions and tasks that they take, notifications that they receive, and the like.
    • Predictive Analytics show us what could happen. Use past and current (real-time) behavior, machine learning techniques and statistical algorithms to identify/predict the likelihood of future outcomes, trends, and patterns.
    • Predictive modeling is a common statistical technique that uses data modeling to predict future outcomes. There are many types of predictive modeling techniques including neural networks, decision trees, linear regressions, time series, etc.
    • Prescriptive Analytics tell us what should happen in the future. In other words, what we should do next. Prescriptive Analytics use data (historical and current) to determine the best course of action/next steps. Marketers use them to help make immediate and long-term decisions.
    • Programmatic Advertising is the use of automated technology for media buying. It allows marketers to buy ads targeting the right viewers at just the right times on a variety of sites.
    • Random Forests are machine learning algorithms made of decision trees. They’re commonly used for regression and classification and can handle huge datasets. Many believe that they provide the best/highest accuracy of all the classification methods.
    • Reinforcement Learning is a subcategory of artificial intelligence and machine learning. It trains the models to make a sequence of trial-and-error decisions that maximize reward/value. Think Pavlov's dogs.
    • Washington Post’s Heliograf, Bloomberg’s Cyborg, and Forbes’ Bertie are all examples of the new machine learning tools that help newsrooms. You can read more about them here. 
    • Set of manually determined rules that predict the sentiment of a social review, blog, mention, etc. Rules-based sentiment analysis is different from Automatic sentiment analysis in that Automatic relies on machine learning techniques/algorithms to understand the sentiment.
    • Say-through rate measures your verbal engagement with voice ads. It’s like a clickthrough rate for voice ads. Some marketers recommend you only measure positive responses (aka affirmations) when you’re calculating your say-through rates. This is short-sighted. Measuring both separately – as(...)
    • Structured data that helps search engines better understand your website.
    • Semantic Analysis understands words within a context, not just on their own. This allows it to get to the intended (aka Proper) meaning of the sentence. So, if you say “I’m beat” it knows that you're tired and not whipped like an egg.
    • Semi-Supervised Learning is a subcategory of artificial intelligence and machine learning that uses a small amount of labeled data and oodles of unlabeled data to train algorithms.
    • Sonic branding refers to the sound of your brand. It includes everything from the sounds your listeners hear when they complete an order on your site to the voice they hear in an interactive ad on a smart speaker.
    • Speech recognition allows a computer to identify words spoken aloud and convert them into text.
    • Google’s Project Starline is a video communication method that allows you to see a 3D model of the person you are communicating with as if they’re in the same room with you.
    • Strong AI (aka True AI) is intended to think on its own. Systems are built with the human brain as the archetype. Makes decisions on/after analysis. Learns. Adapts. Will be better tomorrow than it is today. (Hopefully.)
    • Highly organized and clearly labelled data in a standardized format that is easily accessed by humans and software/programs. Think: Google Docs or Excel Spreadsheets with thousands or millions of rows each clearly and concisely labeled.
    • Supervised learning is a subcategory of artificial intelligence and machine learning that uses labeled datasets to train algorithms.
    • Talking Avatar is an AR technique that allows live interaction between a 3D character controlled by actor in real time.
    • TensorFlow is an end-to-end opensource platform for machine learning.
    • Text analysis uses machine learning to automagically process and examine unstructured text data to find valuable facts, insights, trends, and themes.
    • Using machine learning and natural language processing, text clustering automatically groups text-based documents into clusters based on their similarity. It’s often used to classify products and improve content/recommendations.
    • How fast can the user get the information they’re looking for?
    • How fast can the user buy from you? TTT is the time it takes a visitor to complete their transaction.
    • Tokenization divides text into small clumps of words (aka tokens.) The tokens help The Machine better understand the context.
    • Developed by Alan Turing in 1950 (originally called the Waiting Game), the Turing Test is a game/test where a machine tries to fool a human into thinking it’s a human.
    • The uneasy feeling we get when robots and virtual agents are too human-like. Sophisticated CGI animation in video games and virtual agents with human-sounding voices are both examples.
    • Unstructured data does not have a predefined model. It’s typically scored in its original/native format. Think audio, video, PDFs, text messages, call center transcripts, social media posts, survey responses, and the like. Because it’s not organized in a specific manner, you’ll need things(...)
    • A machine learning technique in which the models are not supervised and instead allowed to discover patterns on their own.
    • What a user says. An utterance is different than Intent. Intent is what a user actually means when they say/write something.
    • Vectorization has several different definitions. In the context of this site, vectorization is a technique for converting text into vectors of real numbers.
    • A computer-generated environment with scenes, objects, and beings that appear to be real. Your VR experience is simulated to be just like the real world or completely different from it.  
    • Voice-added is an option for assisting with input. Voice to text on your smartphone, for example.
    • Voice Biometrics use a person’s voice to verify their identity. (Like a fingerprint but using your voice.) Marketers use them to hyperpersonalize user interactions/journeys. Many other companies use them for authentication and security. Voice Biometrics are used in all sorts of telephone(...)
    • Marketers commonly refer to Voice Recognition in two ways: Voice recognition allows a computer to identify words spoken aloud and convert them into text. (This is actually speech recognition.) Voice recognition allows computers to identify a person based on their individual (unique)(...)
    • Trigger words/phrases that activate smart devices.
    • Weak AI (aka Narrow AI) is a type of AI that is limited to a specific/narrow area. It relies on algorithms and programmatic responses to simulate human intelligence, generally with a focus on a specific task. Example: When you use Alexa to turn on your lights.
    • Works with Alexa Devices are those you can control using the Alexa voice service.  Things like smart thermostats, plugs, and lightbulbs.

    Sign up for our Newsletter
    Learn about the latest Internet marketing techniques from one of the foremost experts in the field.

    You’ll be entertained and enriched as Amy presents proven online marketing techniques with her signature wit and engaging sense of humor.

    Sign Up Now!

    Double your Marketing Power

    Catalogers and other offline marketers have many advantages when it comes to E-Commerce.

    Unfortunately, most of them often turn their positives into BIG negatives. We’ll help you leverage the benefits of this valuable customer segment.

    learn more

    Our Latest Articles

    Eight by Eight  110 Kimball Avenue, Suite 119, South Burlington, VT 05403   info@eightbyeight.com   Privacy Statement  Terms & Conditions   ©2018 - 2024 Eight by Eight, LLC
    Start Increasing Sales Now!

    Get Our Free Newsletter