Artificial Intelligence (AI) as a phrase is bandied about to refer to any number of technologies currently in use. And it’s not that this is wrong per se, but it’s like referring to rustic Italian cuisine and molecular gastronomy simply as “food”. The world would be a poorer place without either, but they serve entirely separate purposes for the palate. According to Gartner, “By 2025, proactive (outbound) customer engagement interactions will outnumber reactive (inbound) customer engagement interactions.”
The distinction being made here is the AI as it is designed for use in the reactive realm (think chatbots) vs. the use case of proactive engagement. While the core technology that underlies both may be similar, and both have specific use cases, proactive engagement is a more focused utilisation.
If you have ever attempted to play the game ‘Twenty Questions’, you have had an inkling of what a chatbot is attempting to do, i.e., asking a series of questions of an individual in an effort to get at an answer. Except in the case of chatbots, you are usually playing the game with an irate customer in a negative frame of mind. Proactive engagement on the other hand is reaching out to the person with a clear objective, ideally before they have a chance to become irritable.
Given that the technology underlying both proactive and reactive engagement is similar, what are the points of difference between the two?
AI startup founders reveal their artificial intelligence trends for 2021
Differentiation between proactive and reactive technologies
| Proactive Communication | Reactive Chatbots | |
|---|---|---|
| Use case: | Knowing the objective of a conversation allows for proactive engagement with a customer. The conversation can be tailored to a specific topic and the language used can steer the customer towards likely keywords (e.g., appointment, bill shock etc.). It also means that the company can control the narrative over a long period of time, which allows for a higher degree of specialisation. | This is in contrast with having an open channel of communication for a customer to reach out to you (usually in case reaching out by telephone and speaking to an agent is not viable). This should exist as a complement to proactive engagement to ensure that a customer can engage with the company for simpler cases. |
| The number of intents: | Focused reasons for communication mean that only the most necessary intents need to be trained for proactive communication. The dialogue can steer customers towards keywords and stay within certain boundaries. | Since a chatbot must effectively operate as a funnel (starting broad and tapering to the specific intent), it needs to have the ability to parse any number of questions a customer could ask in any number of ways. 40-60 intents are not uncommon for the average chatbot today. |
| Less time consuming/expensive: | As a logical extension of the above, proactive engagement requires fewer intents, hence less labelled data, less time to train, and less expensive to train. | Since a chatbot has to be a jack of all trades, it is more time intensive to train (following the logic above). Additionally, add this to the fact that only 9% of customers fully resolve their issue in a reactive realm, which by inference means that inbound call volume is not reduced by a huge amount, which leads to additional costs. And when you add in language variants (to account for Spanish for example), it would double. |
| Length of conversation/follow-up: | With proactive engagement, you are aware of the customer’s contact details and can reach out to them not only at the start of the conversation, but also for follow-ups to see whether an issue has been resolved. While customer journey specialists will be the ultimate designers of fencing in “over-communication”, research suggests that following up on an issue is valued by customers and helps in customer loyalty/retention metrics. | In the reactive realm, historically, customers rarely entered contact details prior to starting a conversation and thus, following-up with them is more problematic. Some chatbots are starting to collect contact details at the start of the conversation, but the follow-up on those conversations is still minimal. |
| Psychological context: | When reaching out proactively, the customer is usually in a neutral state of mind. This neutral state allows you to frame the expectations that customers have of the specific interaction. If you start a dialogue with a customer saying “this is about your appointment on…” they are not expecting help with billing, service issues, equipment, etc. | This is in contrast to a predominantly negative state of mind when a customer starts a conversation with a chatbot or on the phone. When customers initially engage, they predominantly have a problem and have a set expectation where they would ideally want to have their issue(s) resolved, no matter what the topic is at the onset. And since the conversation may not immediately be geared towards keywords (think back to the 20 questions analogy), it takes longer for the reactive chatbot to arrive at the crux of the customer’s problem, while the customer may be becoming increasingly agitated at the pace of progress. |
The prime takeaway from the above comparison is that while it reads more like a pro/con list as opposed to a comparison, these are very much complementary technologies as it is impossible to be proactive for every question/concern a customer may have. In those times, it is imperative to have a channel of communication via which the customer can engage. Hence, while it may be true that 70% of customers have a more favourable view of companies that engage proactively, it does not take away from the fact that reactive channels are necessary as well. That said, the added benefits of cost savings (via reduced calls) and/or added revenue (via customer loyalty and upsell) make it altogether unanimous that proactive engagement needs to be part of your customer experience suite.
