PRODUCT MANAGERS ARE FALLING BEHIND
A 2017 study by DBR Research from StrategyAnd about who is managing AI implementations in financial institutions, which shows that CIO/CTOs, Marketing, and Strategy are leading the charge while product management is leading the AI strategy in only 26% of organizations. (See Chart I). While I cannot statistically extrapolate this data across all industries, anecdotally, we are seeing similar proportions in our conversations about chatbots. Technology, marketing and strategy departments are leading the charge: embracing new technologies in artificial intelligence and regularly experimenting with them. These are also the teams that seem to really understand the value of chatbots and their possibilities.
Moreover, people who want to use artificial intelligence technologies appear to have less in common in terms of job sector and role, and more often fit a common personality profile: a passion for technology, a thirst for new knowledge in their sector, inventive, and have a drive to find solutions to business problems for which their current toolsets are not adequate. And while these personality types seem to describe product managers to a “T”, we generally find conversations with product managers about chatbots to be somewhat fruitless and uninteresting. 9 times out of 10, after a discussion about chatbots, product managers will look bored and blinking, almost always saying, “Yes, but we don’t need chatbots; we can solve that same problem with intuitive design.” (As if it is an offense to their skills that this technology would even be considered.)
INTUITIVE DESIGN v. CHATBOTS
Truth be told, there are many executions of chatbots that are just band-aids for good and intuitive design. One of the most common use cases for chatbots is information flow. Prospective clients often come to us because chatbots bring attention to specific information that is currently being lost on a website or mobile app. For example, we spoke with a prospect for which 20% of customer service calls can be attributed to people not being able to find a location to sign-up for a class near them. The first problem is that the class location is in a sub-header that’s not easy to find. The second being that the tool to find class locations is somewhat difficult to use for those who are, um, technically-challenged (though I imagine any millennial would not so much as blink an eye—it’s quite easy.)
And the right decision in a product manager’s eyes would be to redesign the site and page to provide this information in a clean and easy-to-use way. Rationally speaking, I would agree with this choice. Create a cleaner homepage and an intuitive class location tool, and the resources you require to execute on this would likely be much less than the costs incurred by the customer service team for 10% of their day over a period of months. So in this way, I understand the product managers’ beef. Why use chatbots? Why not fix what we have, instead of adding a tech band-aid. Don’t put lipstick on the pig.
However, there are other, more practical considerations. Firstly, anyone who has been at a company and been through a website re-design will understand that a company website is not done in a vacuum. For many companies, especially larger enterprises, this exercise can end up being a somewhat politicized endeavor where images and content can be a battleground between different departments; redoing the site often opens up opportunities for technical errors, re-working processes, etc. Hence, building or redoing a new website or parts of a website, is no small task; it is one for which great thought and time should be taken. And, it’s often better done, not in parts, but in a holistic way. Sometimes a chatbot can serve to supplement current designs and gain a better understanding of what information a user is missing so that when the timing is right, and a website is revamped, UI and UX choices are made based on data (easily gathered from a bot).
All this aside, there are things a chatbot can do quite swiftly that even the best and most intuitively-designed website or mobile app cannot. And these are the opportunities that I think the product management community is often over-looking and under-utilizing.
WHAT BOTS CAN DO, THAT GOOD UI/UX CAN’T
1.) Targeted Messaging Based on User Demographics + Preferences
Good and intuitive design does not always adapt to user demographics and preferences, but chatbots do. Chatbots can be triggered based on a URL, so as a user comes to your website (or through a deep link to your app) you can trigger a particular bot that can continue the conversation. For example, say you, dear reader, after reading this blog post want to go to our website. We can create a message that continues the conversation.
Once you go to our website, you will see a chatbots that will confirm that you came to this blog post and can also gather other information (e.g. are you a product manager, have you used bots before, etc.) This will help a.) continue the messaging that intrigued you in the first place, b.) gather a richer profile of an anonymous user from our blog post, so we know the demographic audience of our blog post, and c.) if you come back we can have a different conversation, knowing that you have already been to our site.
Good design may make your website or app intuitive, but chatbots make your website or app personalized and hence, more useful.
2.) Dynamic Questions + Lead/Information Gathering
Chatbots are also dynamic. They can easily respond to the user’s behavior and adapt questions or responses based on this information. Take, for example, something as simple as a contact form. No matter how well-designed, contact forms rarely adapt to a user’s preferences and demographics. You can use more expensive and comprehensive systems such as HubSpot or Marketo to do this, but it’s sort of like using a Mack truck to kill a fly. And you can create a system for doing this on your back-end, but editing and updating it will be a royal pain and may often take a cross-functional effort in many companies.
Chatbots are designed to adapt on the fly to different demographics of your user base in a quick and easy way. So pretend you’re a digital agency and you get inquiries from both potential clients as well as freelance designers and writers. Instead of having a basic ‘Contact Us’ form, a chatbot can have a conversation asking why they came to your site. For potential clients, you might want to ask information such as “sector”, “how you found us”, etc. For potential freelancers, you may want to ask them to upload their portfolio, or ask what client work they are most proud of. And chatbots will remember these profiles, so that as people come back they are curating conversations in a different way. For example, if a potential client in the pharmaceutical industry comes back to the website a second time, maybe a chatbots might say, “Did you see our new client success story in the pharma space?” (Something that might not make sense to have on the website. This is not something your average contact form/web page will do.)
3.) Validating Product Choices + Training Users on New Features
One of the goals of a product team is to have product decisions validated; to gain feedback on a features and tools so that they can be improved upon; and to train users on new features. Right now this is often done by user focus groups, or user testing; but it cannot be applied systematically across groups of users automatically. Chatbots, however, can be embedded within a product or platform to easily “talk” to users about how to use new features, and to gather direct feedback as they are in the process of using it. And unlike feedback systems or surveys that are innately built into products, chatbots understand who they are talking to (based on collected demographics, preferences, location, behavior, etc.) and therefore you can talk to segments of users.
So here is the solve. Product managers can start testing these assumptions. There are several really good chatbots platforms that allow you to test chatbots with very little resources (like ours) on certain pages, or even something as simple as a microsite. And the truth is: you need bots, and bots, dear product managers, need you.