7 Reasons Why Chatbots Will Fail

Are chatbots “inevitable” and is their proliferation something we can take for granted? It’s easy to get carried away and to think that their success is a foregone conclusion. Standing in the way of the rise of the chatbots, however, I can see at least 7 major hurdles that will need to be addressed.

  1. AI is not quite ready for primetime.

Every year we hear that AI and NLP are on the cusp of changing everything in our lives. Every year we get our hopes dashed (probably as a result of overblown expectations). Yet every year, we also hear that this time it’s different. Well, I believe this may in fact be true but we are still probably 1–2 years away from seeing commercial AI that really delivers a seamless user experience.

Solution: Be patient. Don’t over-sell AI capabilities right out of the gate. We’ll get there soon enough but we may need more human involvement in the early phases. At least until you’ve accumulated enough user inputs to expand your knowledge library and develop more contextually-aware AI. The sooner you start collecting user inputs, the sooner you can start automating a growing majority of interactions. AI is very much the “secret sauce” but its still getting perfected in the kitchen.

2. Competing Messaging platforms will fragment the market.

Between WhatsApp, Messenger, Kik, WeChat and others, we are seeing several massive messaging platforms emerge creating what’s been referred to as “walled gardens.” There also appears to be a very strong geographical and demographic element to several of these. Will the fragmentation of messaging ecosystems undermine the growth and potential of chatbots?

Solution: No easy fix here. Needless to say, successful chatbots will need to bridge the divide between the world’s major platforms. Sounds easier than it probably is. Nevertheless, chatbot developers will need to envision a future where users can continue a conversation even as they switch from one platform to the other. If the goal is for seamless and frictionless conversations, then shouldn’t every chatbot be platform agnostic?

3. Over-hyping the chatbot “revolution.”

As exciting as chatbots are, there is a natural tendency for “insiders” to proclaim a new era while raising expectations above what’s reasonable. We’ve seen this before in many industries and chatbots are probably no different. Furthermore, we don’t want to create “bot fatigue” even before chatbots have gotten any mainstream traction.

Solution: We need to temper our exuberance and not fall prey to excessive or over-zealous optimism (Dot com fever anyone? Tulip-mania?). Chatbot articles and discussion forums need to stop proclaiming that the technology will transform the world. We all need to bring a minimum amount of pragmatism to temper the frenzy lest things get too frothy. Already, I fear that the media enthusiasm for chatbots may getting ahead of itself and be setting us up for a first major round of dashed expectations. Perhaps its inevitable but until it happens, we should all bring a dash of self-restraint or at least impartiality to chatbot discussions.

4. Drowning in (bad) bots.

Because chatbots are fairly easy and cost-effective to build, we might be setting ourselves up for a world with too many chatbots doing pretty much the same thing. This is already what’s happened with Apps (over 2 million apps for iOS and another 2 million on Android). How will users find their way through so many bots competing for their attention? Worse yet, are we also setting them up for a constant barrage of unsolicited (nuisance) messages from mediocre bots?

Solution: Like with most things, the crème rises to the top. Focus on things like quality content, exceptional user experience, policy transparency, and keep the users foremost in mind. Consumers will quickly sort the wheat from the chaff. The slightest abuse of the “social contract” between bots and their audience will cause the latter to quickly shut down the conversation. So tread lightly and try to be as “frictionless” and non-intrusive as possible. Oh, and being one of the first-to-market won’t hurt either.

5. Data security and confidentiality.

This one can hardly be surprising. The more we interact with chatbots, the greater the concern over protecting our personal data and the conversation history. Hacked or even hijacked conversations pose a major risk to the adoption of chatbots.

Solution: Given that these are still early days, chatbot developers and the platforms they use, should proactively do everything they can to appease security concerns. With things like credit card information and personal health data being exchanged, we have to take every possible measure to protect the trust of chatbot users. How this works exactly, who knows? What is certain is that every chatbot will need to have the very best security “baked in.” Healthcare may be a good place to look for examples and standards for high-level data security (e.g. HIPPA).

6. Generational appeal.

It’s easy to think that chatbots have a stronger appeal with younger audiences (think Millenials). Early data clearly shows that the younger you are, the more likely you are to use messaging platforms and chatbots on a daily basis. The danger here is that chatbots could get relegated to being a tool/platform designed primarily for younger users. This could potentially leave older users (e.g. Baby Boomers) marginalized or sidelined. For chatbots to succeed, they need to be designed to work for EVERYONE.

Solution: Chatbots should be designed and tested with a variety of users, including older ones. Especially for health & wellness applications, seniors will need to become comfortable using chatbots. This may have significant implications for things like content, tone, UI/UX, and things like support/assistance.

7. Monetization & Revenue Models.

The proliferation of chatbots, especially ones that provide a service or support (not just to prop a brand or sales), begs the question of how they will be paid for and by whom. Just as apps have struggled with finding the right business model (Pay-per-use? One-time upfront payment? Freemium? In-app purchases? Subscription models?), so will chatbots struggle to show how they can generate value which can be monetized and easily paid for.

Solution: Experiment. Use trial and error. Poll your user base. Dig through the free-text entries to identify any issues, ideas or even positive feedback that you can capitalize on. Regularly ask your users for quick feedback. Was I helpful? What else can I do for you? What can I do better? Would you recommend/refer me to your friends? Track daily/weekly interactions and the number of conversation exchanges to assess the amount/level of user engagement. Is it sustained/able? Collect testimonials. The great thing about chatbots is that you already have a built-in sampling/polling audience and can collect feedback almost instantly.

These are only a few potential hurdles. While there are probably several others, I remain bullish on the future of chatbots. Especially given how they can significantly improve patient engagement for health and wellness applications.

At Medtech Mojo, we are developing a Healthcare-dedicated patient engagement platform that is based on the very latest Chatbot technologies and capabilities. With our combined 45+ years of Medtech & Pharma experience, we are creating a digital health platform that enables anyone across the Healthcare spectrum to quickly and easily create their own Chatbot. Zero programming required. To learn more and to register as one of the early Beta users, just visit MojoBots.

Andrew Hyncik is an International Strategic Marketing & Product Development veteran with 20+ years of experience in the MedTech, Healthcare and Pharma industries. For more Medtech insights and original articles, visit Medtech Mojo Better yet, try our very own Medtech Mojo chatbot here.