Skip to content

Chatbots, AI, and Context: Top Takeaways From Intelligent Content Conference

chatbots-ai-context-icc-2018-takeaways

“Technology is outpacing our ability to comprehend what we can do with it.”

These words from this year’s Intelligent Content Conference emcee Robert Rose, CMI’s chief strategy advisor, no doubt resonate if you’ve been in marketing for more than a week.

With so much happening so fast in the realm of tools and strategy, what’s a marketer to do?

Marketers learned firsthand at ICC from content professionals who are experimenting with the latest tools and strategies and have lived to tell their tales.

As always, I’m hard-pressed to squish the richness of this conference into a blog-sized summary, but I’ll take a shot by highlighting three main takeaways:

  • Chatbots may be the straw that finally forces marketers to break the copy-paste habit.
  • Artificial intelligence exists today that can automate and enhance many aspects of marketing
  • Context – enabled by a content strategy that transcends departmental silos and funnels – makes the customer experience.

(Thanks go out to the CMI editorial team for the extra eyes, ears, and gray matter between the ears that made it possible to pull together this post.)

Chatbots may break our copy-paste habit

Since ICC emphasizes content technology and strategy for marketers, you won’t find it surprising that speakers and attendees did lots of chatting about chatbots.

What are chatbots?

Chatbots, as noted in this recent post, are “apps within apps” – tools that automate conversations between humans and computers. Chatbots, aka bots, enable scalable one-to-one, natural-language interactions (sometimes text, sometimes voice). Here’s one simple example of a bot from Taco Bell:

tacobot_preview

As Gavin Austin said in his talk, What the Bot!? How Salesforce Geared Up for Chatbots, “A chatbot is another content channel.”

How to (and not to) prepare content for chatbots

Sharing this view of bots as one of many content channels, Cruce Saunders made the case that marketers developing chatbots should avoid the temptation to create a whole new batch of (largely duplicated) content.

Rather – as Cruce said in his talk Unifying Multichannel Publishing With a Master Content Model at Mayo Clinic and has been saying for some time – companies need to develop content models that enable existing content to be reused from a shared repository, not copied and pasted from one repository to another.

Companies need to develop models that enable existing #content to be reused from a shared repository. @mrcruce Click To Tweet

It makes business sense to reuse answers to customer questions from a unified set of content modules delivered across channels: chatbots, product documentation, training, marketing materials, and so on. Content reuse reduces costs of development, review, and maintenance, and increases consistency and quality.

Noz Urbina reinforced this theme (in his talk Chatbots: How They Can Be Integrated into Your Existing Content Strategy) when he said that, ideally, you pair existing content with bot questions instead of wasting time writing fresh content for a bot. For this kind of single sourcing, he says, you need intelligent content.

Pair existing #content with bot questions instead of wasting time writing fresh content for a bot. @nozurbina Click To Tweet

In other words, you need structured (topic-based) content that is chunked and tagged according to content types.

Content in context for chatbots and more

As content strategist Carrie Hane says, “Structure sets your content free … it enables the same bit of content to be delivered to your wrist (via a smartwatch), to your living room (via a chatbot like Alexa, Amazon Echo’s voice-driven assistant), or to anyplace else you choose to receive information via a nearly infinite number of channels and devices.”

Structure sets your content free, says @CarrieHD. #intelcontent Read more >> Click To Tweet

When chatbot content is intelligent – that is, well-modeled and well-tagged – you can personalize its delivery. “Different industries or audiences can get served different chatbot responses … guided by the categories applied to the content chunks,” Noz says. For example, a chatbot might point one audience to a marketing case study and point another audience to a case study on user documentation.

That example from Noz was one of my favorites because it upholds ICC’s long-time vision of content professionals across an enterprise coming together to approach content as a shared asset, with all departments using it in a unified, strategic way.

Here’s the good news: Setting up content for success with a bot translates to success wherever that content is used. As Noz puts it: “What you must have for chatbots is extremely helpful everywhere else.”

Noz defines chatbot-ready content as high-value content that has been tagged according to a taxonomy and structured for intelligent (single sourced) use across channels.

noz-chatbot-ready-content-example

Don’t take on Alexa – at first

By the way, you don’t have to create a chatbot at the level of Siri or Alexa. “Think about what’s unique to your organization and what you can do,” Gavin says. “If you try to build Siri or Alexa, you’re going to fail,” he adds, quoting Salesforce product designer George Hu.

salesforce-product-designer-speech

Start small and simple, Gavin advises. Answer your customers’ most pressing and frequent questions. For example, at Salesforce, where Gavin works, the most-asked question from customers is “How do I reset my password?” One of the first things Salesforce did with its chatbot was to automate answering this question to cut the number of hours customer service reps spent answering it.

Wanna build a bot? Think about what’s unique to your org and what you can do. @GavinAustinSays #intelcontent Click To Tweet

Val Swisher (in her talk Artificial Intelligence: Changing the Game for the Future of Content) similarly assured her audience that not all chatbots must be sophisticated like Siri or Alexa.

“A chatbot is a front end to something,” she said. It could be the front end to any number of types of information system. It could connect to a high-end cognitive system that learns as it goes (including artificial intelligence and natural-language processing). Or, it could connect to a simple branching script that provides a small set of answers to basic questions.

Takeaway: When it comes to chatbots and your company, start by automating answers to your most-asked questions. Treat each bot as simply another channel for the high-payoff content you have, evolving that content for efficient cross-channel use rather than creating unsustainable chatbot-specific repositories.

 

Artificial intelligence gets real for marketers

Pick a task, any task. Something you do every day on the job. Chances are it’s on Paul Roetzer’s ever-growing list of marketing-use cases – 51 and counting – for which some kind of artificial intelligence solution is in the works if not on the market.

AI for marketers is no longer a someday thing.

In Paul’s talk, How to Get Started with Artificial Intelligence in Content Marketing, I could practically hear lightbulbs go off in people’s heads as they reviewed his worksheets and realized many of their familiar marketing tasks map onto existing AI tools, however immature those tools may be.

We’ve covered many of the automatable marketing use cases Paul talked about. What’s particularly interesting is how fast this list is expanding.

For a glimpse at the growing list of AI-enabled use cases and some of the AI tools available for each, start here:

Takeaway: If you’ve been sitting on the AI sidelines thinking its marketing impact is years away, it’s time to think about a game plan for automating or improving repeated tasks.

It’s time to think about a game plan for automating repeated tasks, says @marciarjohnston. #intelcontent Click To Tweet

Context makes the content experience

In kicking off the conference, Robert looked back on 10 years of ICC conferences, all of which he has spoken at. He noted one theme has persisted from year to year: Context matters at least as much as content.

Context matters at least as much as content, says @Robert_Rose. #intelcontent Click To Tweet

This topic – the importance of context – points to ICC’s emphasis on content strategy. Robert urged marketers to consider that a perfect message makes no difference for the business if it’s delivered in the wrong context.

For example, he told the story of violinist Joshua Bell, who played exquisitely at a Washington, D.C., Metro stop during rush hour and was ignored by almost everyone.

Spectacular content plus wrong context equals spectacular ineffectiveness.

Context as strategy

To think strategically about content is to think about all aspects of the experience, including the context. For content professionals, context factors include a mix of the human and the technological, the analog and the digital, as Michele Linn has listed, quoting one-time ICC co-organizer Scott Abel. Context includes:

  • The right person
  • The right content
  • At the right place
  • At the right time
  • In the right format
  • In the right language
  • On the right device

This goal sounds simple and obvious from the audience’s point of view. From the content pro’s point of view, though, it takes a lot to hit this context sweet spot.

Among other things, hitting the context sweet spot requires a unified content strategy, something that ICC founder Ann Rockley has been talking about for decades.

Hitting the context sweet spot requires a unified content strategy, @arockley. #intelcontent Click To Tweet

A unified content strategy cuts across department silos. It accounts for front-end and back-end issues. It requires setting up a coherent system of metadata that enables machines to do the right thing with each chunk of content automatically.

In fact, the way content might need to adapt to a given instance of use is determined by more than a dozen context factors, such as these:

  • Device (operating system, mobile, tablet, desktop, screen resolution)
  • Locale (time, location, velocity, humidity, temperature)
  • Person (age, gender, stage of life, language, relationships)

When companies fail to manage content intelligently to accommodate a variety of factors in their audiences’ context, all those pieces of content – those expensive business assets – may go underused and underappreciated, like heavenly music played to unheeding commuters.

Takeaway: Don’t be Joshua Bell, delivering fabulous content in a context in which people won’t get value from it. Think strategically about all aspects of the content experience, using technology behind the scenes to support your strategy.

Conclusion

Marketing is approaching a point where the old ways of creating, delivering, and managing content won’t work. In his ICC content experience workshop, Noz called this point “the channel experience singularity.”

Here’s how Noz defines this kind of singularity – as the point at which: “Channel proliferation accelerates beyond a brand’s ability to have channel-specific teams, strategies, and structures and must approach experience as an omnichannel continuum, supported by an ecosystem of touchpoints.”

We’re in the midst of a massive expansion of communication channels and formats. Smart speakers. Virtual and augmented reality experiences. Chatbots. Whatever comes next.

As channels proliferate, the only way businesses can keep up with the customer-driven need for omnichannel experiences is to do this challenging work. Knuckle down and create the content models. Figure out your taxonomies. Tag content strategically.

Doing the work to create truly intelligent content supports the channels we know about today. And it sets us up to meet whatever comes next.

In the end, your audiences – and, therefore our businesses – will thank you for it.

Want to gather with thousands of marketers in the ideal context? Register today to attend Content Marketing World Sept. 4-7 in Cleveland, Ohio. Use code BLOG100 to save $100.

Cover image by Joseph Kalinowski/Content Marketing Institute