Closing Panel Discussion: The Future of Content

5 year ago, chatbots, and then nothing. And then metaverse, and then nothing. Now, chatbots. So how to prepare for the next 6 months. 

Self-help is becoming more important. Less about how to document, and more about in-product through onboarding. See confusion coming down the pike. So many possibilities. Will be a case of haves and have-nots. People who are working in environments where they have to use Word are not valued. Google search results have begun the process of enshittification, sponsorship has made it worse. Everyone running to ChatGPT because it tells you things. Those things aren't that good, but it feels better. Vendors need to start implementing AI to serve both authors and end users. 

Voice search

Ultimately going to be another modality. Going to be preeminent as we learn how to master it. If we get our content ready for it, if we plan for it, such that we untangle written words, that is the north star. Lot of work to do to understand how we architect for it. Wonder if we won't see a change of everyone thinking that they need to go to Google first. The cost of creating bad content is trending to zero--which appeals to a lot of companies. It's going to be really critical to be able to find the not-junk. Need some kind of trust mechanism/certificate method to identify trusted content. 

How seeing the integration of TechComm and MarComm?

Content-as-a-service. Still a lot to do to make that work for us. Sources still need to rely on content. The data is only as good as the data is. Not a technology problem. It's a culture problem. Need to get groups together and ensure that you're saying the same thing the same way. Difficult to get C-levels to buy in to cross-org initiatives. 

Crystal ball for the future of TechComm

Better at using AI to curate content. Blurring of enabling content, melding of silos between content types. Less important who created the content, more important that the customer succeeds on their journey. Power of AI will crystallize content, and learning the importance of precision so we author the right information. The idea of "coolest doc portal ever" will go away because info has to be in the product.

AI Unveiled: A Crash Course for Content Developers

Been hearing about AI a lot, because of the release of ChatGPT. Focus on natural language processing, artificial neural networks, and large language models. AI is more than that though. A lot of this has been behind the scenes for some time. Don't know when it turns into "AI."

Natural language processing provides AI with an understanding of the meaning and intent of content. AI doesn't know about anything; it knows how we talk about it. NLP systems have to content with English. English doesn't stick to its own standards. 

Think of your content as a basket of bread, each loaf is a sentence, each word is a slice. In NLP it is called a "token." NLP can figure out meaning by breaking sentence into token and "somehow" understanding. Really learning patterns of speech, patterns that it is taught, as well as content we give it. 

Artificial neural networks are under the covers. It's the framework for AI computing. Subset of machine learning. Inspired by the neural network of the human brain. Part of AI that's doing all the processing. Content creators don't touch that part. ANNs simulate a human brain. ANNs contain nodes, layers, weight, and bias. This is language-specific. A lot of bias in our content, some of which we are aware of and some of which we are not. 

Machines don't understand truth. They can recognize things and predict things. 

Generative AI and Large Language Models. Generative AI is a branch of AI focused on creating new information. Similar to examples. Looking at patterns. LLM subset of generative AI, focused solely on generating text. A system that knows what it knows, not continually learning, used NLU (natural language understanding), understanding more about language than out computers used to. Produces human-like text, understands jargon, idioms, sentence fragments. 

LLMs use ANNs to process and create contextually accurate human0like text. Don't know what they are talking about, just know how we talk about it. LLMs, do not "know" anything in the same way we do. Discerns patterns in the language that we use to discuss a topic or concept. 

Need well-structured data (content) to train AI well. The more consistent the body of content, the better it will be able to give good answers. If you don't train it, the deep learning sill still happen, but the results will be inaccurate, hallucinations, downright dangerous. 

Clean your content. You must clean your corpus. If you do not, your AI will be unreliable. AI does not know if your content is accurate. If you have conflicting content, AI can hallucinate. AI makes stuff up. Beautifully written, grammatically correct.

A KB Doesn’t Have to Be a KO: How to Build Big with a Tiny Team

KB internally for support. Serves thousands of views, updated dozens of times a week, reduces research time for support personnel by 11%.

Content management: repositories. Not going to get the sophistication of a full-on CMS. Includes templates, project boards, AI, QA checking, review cycles and approvals, collaboration and discussions.

KB is focused, fast, well-fed, managed, templatized, future-proofed. Looking for 2 minutes on a page or less. It is not a junk drawer. End up with FAQs and no one can find what they need and ends up being where information goes to die. 

Inspire compliance, the 5 Cs. Clarity (know what to do), competence (know how to do it), confidence (I can do it), consistency (know what will happen after I do it; develops trust), customary (this is how things are done).

Govern intake streams. The process includes intake, triage, schedule, writer, review, publish, sunset. Make the right thing the easy thing to do. Make it easy, make it possible. 

We talked about metrics. Do I care about top 10? Not really. Care about bottom 10. Could be just clutter. 

Automate all the things. Not CI/CD. But have actions drive what you need to do. 

Maintain reliability by campaigning constantly, watch for trends, prep for disaster, maintain calm control. People will forget about you if you do a good job. 

GitHub includes project tracking. 

Reacji action in Slack, attached to emoji, if that emoji is used as a reaction, that gets piped into the doc channel.

Content and Silos and Users, Oh My! Transforming Our Customers’ Content Experience

Medidata had a fragmented content experience. Users could be sent to a support site, a knowledge site, or their own company's information. There was duplicate content, there was missing content. 

Did research to what users expect from a modern customer content experience. They use Apple and Google and know what a good experience is like. 

Want content to be pre-emptive. Get information on screen before they know they need help. When searching, they expect it to be easily findable. Resources should follow a self-service model. Users want to be able to help themselves. User expect consistency across all channels. Users want personalized content, based on their roles. 

A CDP platform by Zoomin, is what was adopted. Applied a unified taxonomy and served it wherever it was needed. Did not need to disrupt authoring. We revamped delivery. A single knowledge hub is the container for all content, and then through the CDP, delivery is accomplished. 

Taxonomy is key. Creates a meaningful organization for all users.  

Best approach to anchoring a strong community of learning is iterative. Build on common assets. Start with structure, then align, upgrade, and deliver. 

Share wins across the organization. 

Continuous improvements through data insights. Use content analytics and user journey insights to evaluate and improve. Enhancing with ChatGPT. Tracking user success within the knowledge hub. 

Use Confluence for authoring, only writers and reviewers can access, along with k51t add-ons. Supported existing translation workflows into 4 languages.

Building Business Value Through User-Centric Content Design

We need to amplify the value that our content creates. If you are able to quantify with authentic metrics, it can make your story more compelling. 

Adopting a "design thinking" approach to content development. It allow sour team to have a "seat at the table" for content development. 

Content Value KPIs. It's a scorecard to quantify value. It includes support cost savings, or support cases deflected by documentation, returning users, revenue growth, and product answers, which increase customer satisfaction.

When entering a support case, gets suggestion of documentation, which can deflect to information where support isn't necessary. 

Other KPIs are increased CSAT scores, a growth in publications, increased SEO, and higher portal traffic. 

How get there? Largely because of adopting design thinking approach, which happened when the manager of technical publications started to report to the director of design.  Data and insights from user research inform and guide content strategy. Get a lot of qualitative data from user research, which has helped realize content outcomes.

Becoming a user advocate helped us become more empathetic with our users.

Design thinking can be somewhat disruptive because you're always iterating.Once you can show the content outcomes and values, that helps to perpetuate the idea.

Headless: Catch Silos If You Can

Back in the day, company had help center, documentation, API reference, e-learning, and certifications. Help center was in one tool, documentation and API reference in another tool and e-learning and certification in a third tool. 

Customers told us when searching, they could not find. They reached out to support, which pointed them to a different system. 

Headless CMS is best explained when compared to traditional CMS. Front end is decoupled from back end. 

Can we even name all the solos? Split into categories: authoring silos and customer silos.

Authoring silos have complicated linking, separated teams, fragmented authoring, and multiplied maintenance. The more tools you have, the more things can break. 

Customer silos have poor discoverability, fragmented user experience, fragmented search, and medium-oriented navigation. 

The solution isn't just to throw everything into a headless CMS. 

Idea to create one education portal. 

Silos can be solved on the team level, but there might still be silos on the company level. 

4 tips to become headless. Realize the impact. (Do the homework.) Plan well. (Don't underestimate thorough planning.) Ensure needed resources. (The bigger the project is, the more people you will need, and from different roles.) Iterate. (Implement one-by-one.)

A Migration Is a Terrible Thing to Waste—A Roadmap for your Next Big Content Migration

What do we mean when we say migrations? When moving data from one system to another, such as a web content management system. 

Want to not just get to the other side, but want everything to be in a better place. 

Migrations are hard & tricky. Take longer than anyone expected. 

Don't waste the migration opportunity. Use it to restructure and improve. 

Important to start by understanding why you're migrating. Could be technical reasons. Could be functional reasons. Could be rebranding. 

Start with a content audit. How many content types? Will tell how many content structures you're dealing with. How many content items. Matters less than you might think. How it is structured: Blobs or atoms? Unstructured or structured? If unstructured, do you want to make it structured as a part of the migration? Any inline styles and other markup sins? Migration is a great opportunity to clean that up. What is your best content? And what can you get rid of?

Can start to think about how you can improve your content even before you move it. 

What can be left behind? Has brand, org, evolved? Is there content that no longer performs well? You can save money by identifying content that you don't need to migrate. 

Identify how your content can be structured differently or better. An opportunity to add metadata.

Stories with Soul: Staying Alive in the era of ChatGPT

 Keith Boyd, of Microsoft, and Mario Juarez, of StoryCo, talked about stories and ChatGPT. 

In 2003, we were talking about structured content, XML, and ebooks. And we would have been talking about AI as well. 

Storytellers set the vision. Steve Jobs said that the most powerful person in the world is the storyteller. 

Storytelling has to be a big part of our toolsets. 

The elements of a good story include premise, plot, character, prose, and theme. Element of humanity in all of these. Stories are always about human beings. Sometimes connected to struggle, which is where you get plot. 

A story is the hero's journey. Not going to need it in tech docs, but it is highest rendition of something wired into all of us. There is a sense of fairness built in to this narrative structure. It resides in anyone who digests your content. 

A lot of problems are global, but starting at that level numbs the audience. Focus on the individual. 

So how to leverage AI? Well, with AI, what could go wrong? Well, AI has a tendency to hallucinate and make things up. 

How to think about AI. It can be a framework to overcome writer's block, how to get started on that first page. It can start with good, and that's how to get to great. In theory, how to get started. 

How to use superpowers to give my content more soul. "SOUL" is an acronym.

Sincere: Your story reflects sincere perspective.

Optimistic: Story inspires positive action

Unique: Perspective is unique.

Lively: Engaging language that positively challenges readers. 

But more: ERA

Empathetic: Connect with audience in an emotional way.

Relatable: Foster meaningful connections.

Authentic: Only you could write this story.

SOUL ERA has several levels. On its best day, content that comes out of ChatGPT is at best at the first level. More or less accurate, but with no soul. 

With craft and care, can get beyond that. Can start with that draft. 

Level 2 is demonstrably different. Uses language such as metaphors that only humans could understand. Level 3 builds a visceral connection, demonstrating deep knowledge, inspires action because the content is actually consumed. 

Generative AI is the ultimate research assistant. You can ask it every stupid question, explore every topic, and it can define every term, validate every assumption, generate sample text blocks, and provide a reference roadmap. 

AI can't be clear, concise, or compelling.

Monday Keynotes

Reltio's Megan Gilhooly noted that making unpopular policies more readable doesn't make them more popular. 

The problem with being a perfectionist is because it inhibits you from reliably using your data. One type of perfectionist looks for the perfect piece of data, never finds it, and never moves forward. Another type thinks their work is more "perfect," and doesn't trust others' work. And another type is hyper-critical of imperfection. 

How to break free from perfectionism when it comes to data. Perfect costs too much. Accept iterative progress. More precise might not be better. 

Maura Moran, of Contiem, addressed getting business buy-in. Trying to make a case for content investment is intrinsically hard. It's a bit niche, people don't understand what you're trying to do, and you're always competing with other projects. 

Key is to make you value visible in a way that resonates to stakeholders.

Identify the real value to your business. Two business case benefits: financial (ROI) and non-financial.

Align the project to company goals and culture. Reflect the language of the organization. 

You have to make the problem visible. May be that people don't think there is an issue. 

Know your stakeholders. Not only people in your workflow. Look for influential allies. Understand the approval process. Need to know when budgets are signed. Know who is on the approvals committee. Have private meetings and get questions answered early. Find a sponsor with influence. 

Have to be constant and consistent and memorable with communication. Your initiative can be cut after it's been approved. One team came up with the tagline "Taxonomies are our superpower." It helped to keep the visibility of their project high. 

Create a brand. Even come up with a logo--and use it!

Samantha Azzarello, of J.P. Morgan, told a story of building a content strategy from scratch. 

Came from an economist background, did not know much about content. Had lots of questions. But I knew if I built trust and credibility, could be successful. 

Content is super-regulated. But also has to be quick, responding to events. 

J.P. Morgan had never done a content strategy. To do strategy would need better organization and distribution. 

Building trust and credibility means building relationships, with honest communication, in the details.

Don't get to do it sometimes. Have to do it all the time, every time. 

Iterate with language, especially if something doesn't resonate. 

Deliverables go a long way to building trust and credibility. 

Being helpful, responsive, collaborative goes a long way. If  a stakeholder asks, if you can help and respond, it goes a long way. Being helpful and responsive can go a long way for when you need to ask for something. Note that this is not transactional.

Content Quality is not an Accident – Defining and measuring KPIs for Enterprise Content Quality

 How do you define quality. Well, take burgers. You can define quality if the size and shape and components are the same. Or you can look at taste. How do you measure that.

One way the the quality of content has been measured in in the readability of the text. 

Need something else. readability scores can be high on unintelligible text. 

What does in mean for linguistic quality? Controlled language. Rules for style, grammar, and spelling, plus a vocabulary. 

A style guide is a formal description of a corporate language. The settings of a style guide are the KPIs. The settings of a style guide define the quality criteria.