AI Assistants vs Chatbots Round 2 🥊

Ranbeer
AIAutomation
Published in
8 min readAug 16, 2019

In case you missed the first round of our AI Assistants vs Chatbots blog series, here’s a brief recap of the bout. With the rise of artificial intelligence, we have seen the widespread use of two automated conversation tools: AI assistants and chatbots. The differences might not be clear on first glance, but they are vast.

AI Assistants vs. Chatbots Recap

Let’s look at what happened in the first round. On the one hand, we have the chatbots. They can respond to users’ text or voice with programmed responses. Chatbots route users by asking questions, and the routes are limited to what has been programmed. On the other hand, we have the AI assistants. They have had more extensive training and learn on their own; a far more agile tool. AI assistants use Natural Language Processing, so they can learn new intents and tailor responses accordingly while engaging users in more natural conversation.

AI assistants had a clear advantage. They are better able to respond to customer needs and offload mundane tasks from your employees. The results are in: happier customers and more satisfied employees; increased ROI and stronger brand image. If you still aren’t sure that AI assistants are the best solution, let’s see how bots and assistants perform in round 2.

Use Cases

To understand the benefits of AI assistants compared to chatbots, let’s see how they perform going head-to-head across five use cases.

Banking

AI assistants provide large benefits to banking institutions and their customers. Because AI assistants can handle repetitive tasks, banks can ensure employees are focused on the customer relationship. Chatbots can direct customers to the right people and even automate some account services. AI assistants, on the other hand, go quite a bit further in improving banking processes. By engaging customers in natural conversations, they provide better customer service.

With accelerator tools like Microsoft’s Dynamics 365 Banking Accelerator, banking institutions can integrate AI assistants with market standard models so that all of their data is kept in a single dashboard, providing instant access to data-driven solutions.

Banking Personalization

Because AI assistants continue learning, they can personalize in-app and online banking experiences by understanding typical user behaviors and providing users with a clean interface that is designed specifically for their needs. For example, AI assistants can provide updated and personalized wealth management suggestions and automated loan assistance. By linking with historical banking and behavior data, users do not need to enter any information manually, streamlining access to services.

Mortgage Automation

Another key use of AI assistants in banking is mortgage automation. From approval to lending, the entire mortgage process can be handled by in an AI assistant interface. Automated mortgage lending converts traffic to customers, affords paperless processing, ensures and easily verifies compliance and reduces labor costs by decreasing staff, even as you scale up. With personalized user information and easy navigation, automated mortgage lending also improves customer satisfaction.

Rapid, 24/7 Support with AI Assistants for Retail

In 2018, retail sales hit $6 Trillion in the US alone. Globally, sales came in at over $24 Trillion. With such huge figures, steadily growing at that, it is no surprise that losses due to poor customer service are also staggering. Forbes estimates that US companies lose $75 billion annually due to poor customer service. If changes aren’t made, the figure will only grow as retail sales grow.

Thankfully, we know the drivers behind these substantial losses.

· Customers who feel unappreciated.

· Customers who can not speak to a knowledgeable employee.

· Customers who speak with unhelpful employees.

· Customers who have to speak to multiple employees before getting answers.

Now let’s see how chatbots and AI assistants compare when confronting these problems. Chatbots are able to provide some routing and scheduling, as well as provide limited access to information. In short, they can replace knowledgeable employees in limited cases, as well as routing users to the right employees for their questions. Unfortunately, many interactions with chatbots are very similar to speaking with an unhelpful employee, albeit a virtual one. I cannot tell you how many times I’ve received the response, “I didn’t understand that” from chatbots. Frustrating as it is, it is simply part of the design of the bot — they can only handle programmed responses, and they do not learn from additional text information.

AI Assistants Boost Retail Customer Support

AI Assistants on the other hand, can address each of these issues. AI assistants can schedule follow-up communications to ensure that customers feel valued and appreciated — not all communications need to be driven by the user. AI assistants can label and respond to a larger number of intents and provide access to a greater array of relevant content. AI assistants have conversations with users that are far more natural than chatbots. That is because assistants rely Natural Language Processing and learn language like we do, through exposure to language and correction. As such, they are far more flexible conversation partners, lowering the chance that users come up against frustrating responses and non-sequiturs.

AI assistants can not only handle a majority of inquiries themselves, but are also able route to the correct representatives using intelligent intent prediction. This provides those representatives with information from the conversation so that the user doesn’t have to repeat themselves. If the inquiry requires a conversation with a human, AI assistants can sync with company calendars to ensure that users are scheduled with the right person at a time that is most convenient for them. Reminder pings offer users the opportunity to reschedule if their own schedules change. In short, AI assistants are customer-centric conversation partners that leave the customer feeling heard and satisfied.

Expanding and Accelerating Healthcare with AI Assistants

Let’s investigate another use case, one that tends to be full of frustration: healthcare. Going to the ER or even to your primary care physician often requires recounting your ailments on the phone, to a nurse practitioner and then your primary care provider. If your treatment requires a specialist, then you are stuck recounting the problem a couple more times.

Chatbots can be of some help in scheduling an original appointment and getting that information to the clinic or hospital. They can ask you a set of questions and schedule appointments with the right care providers.

Improve Patient Experiences

AI assistants, on the other hand, can ask a series of questions to guide physicians toward diagnosis and even specialist follow-ups. Upon arrival, each interaction is simplified because every practitioner is provided with the necessary patient information so that treatment can happen quickly. After the initial visit, the AI assistant can then provide you with a specialist’s contact information and set up reminders for follow-ups. Remote patient monitoring then keeps up with you post treatment to assess any issues in treatment. For example, if you are prescribed a medication and you are having unpleasant side-effects, your AI assistant can schedule a follow-up appointment to adjust dosage or change medications.

Assist Providers with Automation

The benefits of AI assistants aren’t just for patients, however. Providers can save time and money by employing them effectively. By having the right information on hand, providers can focus on the provider-patient relationship and treatment. If a follow-up appointment is needed, the assistant’s conversations give providers the information they need to make treatment adjustments and have the conversations to keep patients satisfied and healthy. Using AI assistants in tandem with accelerator tools such as Dynamics 365 Healthcare Accelerator provides healthcare professionals with immediate data-driven insights for care innovation.

Government

Public policies seek to improve the lives of residents and meet their needs through representation. The costs of resident participation are often high; many residents have low information access and little time to participate in the political process aside from voting. The highest turnout for a midterm and presidential election hit just over 53% and 58%. The question for policymakers and government bodies ought to be how they can begin to represent and solicit feedback from nearly half of their voting age population.

Fortunately, AI assistants can be a huge help on this front in two ways: by providing intelligent 311 and by soliciting resident feedback on government websites, social media profiles and social media posts. Let’s talk about city services first and then dig into how AI can aid the policy process.

Reduce costs with Automated 311

The average 311 call costs $3.40, a cost that could be reduced with intelligent AI solutions. AI assistants can handle the bulk of 311 calls and webpage visits with automated services such as scheduling bulky item and hazardous waste pick-ups. Because AI assistants learn on their own, automated 311 reduces labor costs, which includes training and retraining. For more complex questions, the AI assistant can direct callers and site visitors to the information they need without human intervention. Follow-up chats can also gauge resident satisfaction with city services to provide officials with up-to-date feedback into sentiments. This information can help officials locate substandard and excellent services in real-time.

Understanding Resident Sentiments with AI Assistants

AI assistants can also be easily integrated into government websites and call lines so that links can be embedded into social media posts for higher engagement. This gives governments the opportunity to show that their residents’ thoughts matter and provides administrators with the data they need to see how policy roll-outs affect their residents.

If AI assistants are used with social media monitoring, links can automatically be provided based upon post intent, ensuring that policy sentiments can be captured in well-structured data. Location and demographic information from responses can help locate where services are underperforming or underutilized.

Chatbots can only direct residents to the right people and right services. They can also perform some basic automation tasks like service scheduling. But again, they are constrained by the responses that are programmed into them. Additionally, they are not capable of responding to nuanced intent, and are a poor fit for the complex realities of public administration.

Manufacturing

So far, we’ve addressed the role of AI assistants in a public and consumer facing direction, but they can be as effectively used internally. Manufacturing is a machine heavy process that relies on the ongoing performance of production work-processes. An underperforming machine can yield a large drop in output and revenue.

Automated Machine Monitoring

AI assistants can provide the operational status, output level and maintenance recommendations of machines in all parts of the production flow, providing production staff with unprecedented insights into the functionality and efficiency of the production process. AI Assistants help speed up time-to-market and reduce operational costs. With AI assistants, factory managers can provide specific inquiries for which the assistant will then provide real-time data.

Chatbots, on the other hand, cannot provide up-to-date information as responses need to be programmed. In short, they can only provide manufacturing insights that were already known, costing additional programming time to provide information that was already accessible.

Learn More About Intelligent Automation

AI assistants are one of many effective applications of intelligent AI. To learn more about intelligent automation applications, read my recent blogs or visit our website.

Whether you’re looking for a no-code intelligent automation solution yourself or the right information to evangelize and train your workforce for this fourth industrial revolution, taking the time to understand intelligent transformation is an important first step.

Please go to social27.com to learn more.

--

--