Remember the beginning of Indiana Jones: Raiders of the Lost Ark? Indy delving into the Peruvian cave to find the Golden Idol of Fertility and quickly discovering tarantulas crawling all over his guide’s body. Your site is that guide. Google and Facebook spiders crawl it regularly trying to make sense of your content and discover suggestions for your customers. This blog post explores what it means to build a site for bots first and humans second. It will outline the progression of moving towards developing for Artificial Intelligence (A.I.)-First and provide suggestions on how to execute on this new paradigm in development through the use of structured data and adhering to web development best practices.
The Progression of Experiencing Content
Days of the Desktop
In the old days when we made websites for companies, users visited those websites with browsers on desktop computers. No one said Desktop-First as we lived in a Desktop-Only world. And then came along these things in our pockets more powerful than the mainframes that took Apollo 11 to the moon. At first, even with this newfound superpower, most site visitors still browsed sites on their home or office computer. Then people began experimenting with various apps and built habits around a few core ones like Facebook and Amazon. Google became the search engine. Google became a verb.
Emergence of Mobile Browsing
As more and more people interacted with internet content through apps and their mobile browser on their phone or tablet, the term Mobile-First started circulating in web development and design conversations. Mobile-First means making design and development decisions which prioritize the mobile user experience first and then take into consideration the desktop experience. By explicitly committing to mobile, designers choose images and fonts which display well on smaller screen sizes; UX people think through site interactions with touch events and gestures early; and developers factor in slower internet speeds, the need to pull in different assets depending on screen size, etc. Mobile devices had children: Large tablet children, iPad mini children, Android children, iPhone children. Children with lots of different needs and wants: Browser preferences, screen sizes, operating systems.
Rise of the Machines
Now, in addition to the prevalence of mobile, you have the beginning of growing interactions with digital assistants, think Alexa, Google Assistant, Siri, Cortana, etc. We use them to play our music, order goods, make reservations, tell us about the news, manage our calendar, and find directions. They are useful in getting answers to basic questions and are beginning to become more conversational. There are also strides being made to make them sound more and more human. The young users of these tools quickly adapt to voice-enabled IoT interfaces just as millennials before them adapted to the world of mobile.
The move is towards machine learning, artificial intelligence. In the words of Google: the move from search to suggestion.
Users depend on the internet more than ever but in increasingly siloed ways. We go to Google to find the right answer, Amazon to purchase what we want and need, and Facebook to scratch our social connection itch or, increasingly, express our frustrations with the Other. Most people do not interact with very many websites. They interact with the few to get the information they want, view the article or blog post that their friends or influencers they follow are sharing, or purchase those socks for their daughter’s new soccer season. Similarly to how we now experience our music, our internet content is curated by those we follow and link to and the algorithms which recommend it. Gone are the days where we browse a full website let alone specifically visit a website. The big three or four companies drive the majority of what gets looked at and by whom. Think of the sites you go to on a regular basis in your browsing. How often do you search the web to find a company website? More often, you look to find something they are selling, or for their contact information, address, phone number, or hours of operation.
As a business, you want to serve your customers well. You want them to know when you are open for business, how to reach you, and how to find you. If you sell products online, you want to make that process is smooth, seamless, and secure. You want people to be able to find the information about your products or services quickly, cleanly, efficiently. The best way to help your customers find what they are looking for is to embrace Google, Facebook, and Amazon curating content for your customer. Your job is to best aid those companies with your data to facilitate the customer transaction.
A.I.-First means building a website in a way to convey its information accurately and usefully to the artificial intelligence which is crawling the site and prioritizing the A.I. site experience above all others. It means the most critical aspect of your website is not how it looks, how it manages user experience, or how it handles different browser versions, but instead how it articulates the content of your site and the customer interactions you want to become more seamless to the machines reading your website’s code.
How to Execute A.I.-First
There are two levels to this, and one will seem like it is backtracking on what I just said but hang with me. You are going to want to leverage and implement structured data throughout your site and make your site fast, accessible, and secure.
Embrace Structured Data
A.I.-First means rolling out the use of structured data on your website. Structured data is a standardized format for providing information about a page and classifying its content. It explicitly tells the bots out there reading your page what it is finding. Structured data doesn’t appear to the user; it lives in the file that is being read by the web browser or spider crawling your site. We use structured data because, at the most basic level, when a computer reads the HTML of your website, it sees things like there is an image or there is a block of text but it doesn’t recognize the image in the picture or whether the block of text relates to the picture. Structured data allows for the spider reading your site’s HTML to make better sense of what those images and text represent and connect your content to other data that is related.
Structured data also helps train artificial intelligence. For example, if Google is working on training its spider to recognize images of delicious pizza, structured data will tag specific images as “Pizza.” This information provides a constant influx of seed data for the training of the A.I. on what pizza images look like and the various layouts “Pizza” can be displayed in (whole pie, slice, on a plate, angles of the photograph, etc.).
Understand Structured Data Protocols
Open Graph and Schema.org currently represent the most used structured data protocols. They provide the vocabularies which enable Google, Facebook, and the like to grow their knowledge graphs with the interconnected objects and relationships of your web content.
There are specified content types which are used to create rich results for everything from articles to job postings to events. By knowing these structured data protocols and utilizing them in your structured data content, you enable Google to make better sense of that specific content and allow for the sharing of your content to leverage the rich results giving customers a better experience of your content on Google, Facebook, Twitter, etc.
Plan and Test Your Structured Data
To develop A.I.-First requires thinking through your wireframes and content with an eye to what sort of structured data is most pertinent to each component. Structured data is invisible to the person viewing the page content but needs to relate to what is there accurately. Think about where you can leverage structured data actions like making a reservation for an event or booking an appointment which allows for a cleaner user experience for potential customers. Make sure to use the appropriate schema for the content and don’t try to force content to fit a schema as it is frowned upon by Google.
After you have begun development and have sections of your site that contain the structured data you are optimizing for, don’t forget to test it. There are great tools by Google and Facebook to test how your structured data is being viewed by their bots and what sort of rich results will come from it.
Follow Google’s Web Development Best Practices
Building a site for A.I.-First requires adherence to web development best practices. By executing on these best practices your code communicates to the Google spiders that it is a serious site with much time and effort spent developing it. Google is better able to trust its content. With inconsistent adherence to best practices, the Google spiders have more difficulty recognizing your site’s code from other sites that are more amateur or spammy in nature.
Websites which follow web development best practices are also better parsed and understood by spider bots. Accessibility is especially good for this as it is another way of providing clues for the spiders as to what content is about. Part of good accessibility practice is having shortened meaningful descriptive text for things like images, buttons, etc. A person listening to a screen reader doesn’t want to hear a 30-second description for an Order button for a fleece sweatshirt. They want to hear things like - fleece sweatshirt, available sizes input, price, quantity, checkout button so they can get their sweatshirt ordered. When the bot reads this, it becomes much easier to know what that page is about it and recommend that midnight blue fleece sweatshirt to potential customers looking for recommendations from Google.
Google built a great tool called Lighthouse for developers to test a URL for best practices which can be found in the dev console or run via command line through a npm package. When you process your URLs through it, you will identify areas your site needs improvements in speed, accessibility, and security, as well as identify areas your site is already doing well.
What This Means for Your Company
The bottom line is you want to meet customers where they are. You want your company’s website to attract and retain customers, enhance and maintain brand image, and pull in potential new hires. Building your next site or overhauling your existing site for A.I.-First helps achieve these goals. It reaches your audience and enhances their experience by communicating content clearly and strategically in places where users interact with content like Facebook and Google. Google’s knowledge graph becomes better able to make sense of your information and how it fits with suggestions they offer people looking for help. And, as a bonus for prioritizing web development best practices, your site visitors will encounter an optimal brand experience.