With the chatbots creating tremendous opportunities across the industries for a better customer experience, services and revenue generation, more and more businesses are readily investing in it now. As a result of advancements in artificial intelligence, a huge number of businesses are engaging in deploying bots on messaging apps like Facebook, Skype, Slack, Kik and others.
In view of the rising business potential of Chatbots in different domains, VachSoch, a leading platform for interdisciplinary learning and applications, collaborated with various speakers including Net Solutions through Chatbots news and recently launched a series of botcasts.
A botcast was conducted with Abhi Garg, Head of Products, at Net Solutions on “Chatbots for Business”, covering how chatbots can help increase bottom-line profits, how businesses can customize bots and leverage them and where all can bots be fitted well.
Follow the link to hear the audio botcast.
Q: What is a “chatbot” to you?
A: A chatbot is a digital touch point, a conversational interface, which enables rapid information distribution. It opens up a new avenue for businesses to connect with their customers. It removes friction and artificial barriers that get erected when we observe non-office hours, weekends and holidays. Based on what one is after, chatbots have the potential to be Wikipedia on steroids.
— Net Solutions (@netsolutions) December 15, 2016
Q: How do chatbots help in improving productivity and bottom-line profits for eCommerce businesses?
A: Most businesses offer a transactional relationship. This is especially true for e-commerce. You need something, you look at various options online, compare prices, check delivery in your area, finalize a vendor, add the item to your cart, pay for the merchandise, the item is delivered, and everyone moves on.
The next time you want to buy something – you repeat the process. While mobile apps have made the process a wee bit convenient, you still, however, need to go through the rig morale of search and selection process, which is a time hog. Chatbots provide valuable help in this regard. An aggregator chatbot can look up prices and availability across various e-tailers and give you a quick recommendation of which vendor to buy from at the best price, and assist in completion of the transaction without leaving the Bot interface. On the flip side, let’s assume we don’t have the luxury of an aggregator bot, or that a consumer almost always buys from a given vendor.
In such a scenario, the chatbot specific to the vendor can assist in the purchase process and help with queries that would otherwise have a required a human to be inserted into the equation. For example, if you’re looking for a tablet charger and are not sure if the wattage of the chosen product is strong enough to charge your iPad – you could either read through the specs of the product and compare with the specs of the tablet (which most people don’t prefer to do), or speak with a customer service agent (wait in telephone queue, spend time, talk during official business hours etc.).
Both are time consuming and not hugely appealing options. If the vendor has a chatbot, it could do that for you in an instant by running through its behind-the-scenes data bank. No more scratching your head or relying on an agent to do that for you. The other aspect of a B2C relationship is after-sales service and support. Tracking of order, shipment status, filing a grievance etc. all require customer service agents. This is a money burn situation and can be ameliorated by deployment of bots that take away the repetitive, trivial stuff that majority of the customer queries encompass.
— Namrata Nagar (@VachSoch) December 7, 2016
Q: How do Chatbots fair better compared to human interactions in eCommerce?
A: When you converse with a fellow human being, the expectation is that there would be an emotional connect at some level. More often than not, it does not happen. The volume of traffic on the support lines, long work hours, running through the same mundane script leads to emotional detachment at the service agent level. This fuels the frustration for the customer, who has already patiently waited for some time to get to the service agent. This is not to say that this is always the case, but more often than not, business customer support has rapidly become commercialized, outsourced (not necessarily off-shored), and run like clockwork. The customers need to validate their identity, answer questions that appear to be a waste of time and all of this adds to the overhead that is perceived by the customers.
In short, bots provide the following advantages over human interactions:
- No office hours, chat at your convenience, day or night
- They don’t get angry or lose patience.
- They don’t try to shove an up-sell down your throat.
- No wait time for the next service agent to be available.
- Universal interface – you can buy, browse and get service through a single channel.
- They can be programmed to offer promotions and discounts based on time of year or inventory in hand.
- Once NLP and Speech recognition go mainstream, bots would manifest themselves in a different form and take pre-emptive actions that would preclude a customer from doing a lot of repetitive things (the future of commercial AI).
Q:To what extent are Chatbots employable for lead generation and nurturing?
A: To appreciate what bots can do, it is equally important to understand what they cannot do. In their current form, the simplest bots are script based, if-else oriented logic interpreters. The more advanced ones understand some natural language inputs and can add relevance to what’s being asked of them.
For example: to understand ‘running nose’ means a person has common cold vs interpreting it as a nose that is sprinting on the field. All of this is a WIP. As the NLP engines gain more intelligence and are able to parse the input without the need to load it with keywords, Lead Generation and Nurturing will become more relevant. This will improve further as bots gain contextual awareness. You chat with a bot today and if you restart the conversation the next day or sometime in future, the bot needs to know what was last discussed, and where you are in the process. Once we move out of the realm of transactional conversations, bots will add more value to the field.
Bots cannot replace the nuances of human emotion. Do the bots know when the iron is hot and strike? Not at this point. A lot also depends on the dollar value involved. If it’s a matter of selling a DSLR camera, it’s simpler to fix a price band and let the bot run with it. Feed the results of all transactions into Tensor Flow (or similar) and let the bot learn from its own experiences. If we’re talking about selling a SAP Hybris subscription, a lot of tangential details matter such as the profile of potential customer, the sales target and current achievement level, strategic value of the sale etc. – all of these have a subjective component attached to them and need a human override.
Q:How do Chatbots act as enablers and not job re-placers?
A: Bots are a mini-disruption so to speak. They are not as big as creations such as Internet or Cart Wheel. However, they do have a potential to disrupt an industry that has traditionally been human heavy – customer service, and in some sense, sales. Will bots completely rid humans from these service classes? – very unlikely. There would always be the need of a human override, a gentle nudge down the right path or that last bit that closed the deal. Bots, in the short term, would be the assistants that take the drudgery out of the mix for businesses. They would make the call “primed” for real talk once the basics have been taken care of.
Would there be job losses? Yes, very likely. With any new introduction, the ecosystem finds a way to balance itself. This is relevant for small dollar value enterprises – selling flowers, for example. Will they replace the human touch where the service agent knows to add an extra rose or two to the bouquet that was ordered because the caller is super excited about proposing to his fiancé? No. It will be a trade-off. Some very low level jobs would be compromised but in other ways, bots would be level 1 agents and try to close majority of trivial service inquiries at their level.
Q: What kind of eCommerce businesses can make best use of Chatbots and what’s the scope for customization in designing bots?
A: Here are some use cases
Customer Service: The face of a company. The gauge of the confidence the customers place in it. The measure of goodwill a company enjoys. It is one of the most important facets of a company’s operations. It still, however, is the one area that single-handedly ends up bringing a business’s reputation into disrepute. You could have the best product on the block, but without customer service to back it up, it is not going to go too far.
Why Bots? How do they come into the picture? To answer these questions, we have to understand why Customer Service is a tough nut to crack. Some of the pain points are:
- Time it takes to service a query
- Availability of service during off hours and holidays
- Validation, Authentication and Verification rig morale
- Detached outlook that most service representatives project
Bots take these pain points out of the equation. Instant service. Any time. Without frustration or anger.
Hospitality: Some hotels have started using Bots for room service. Place your order through Bots and the services you need get delivered to your room. Planning a tour, get a download of destinations, room availability etc. through a Bot. Want to buy tickets for skiing – let the Bot take care of it. In isolation, these are all great use cases. Put into practice correctly, they remove the friction that humans introduce. Bots understand the language of zero and one. They don’t assume anything. Feed them input, they process it and give you output based on the intelligence that has been endowed to them.
The natural evolution of the use cases would be if you let the Bot do the heavy lifting for logistics and planning. If you want to go on a vacation and are contemplating between destinations, let your helpful Bot do that for you. Just type: I want to take a vacation in May for a week and I have $5K lined up for it. The Bot then waves its wand and comes back with: Okay, Los Angeles seems to be a fine choice at that time of the year. Santa Monica Beach is playing host to Country Music Festival in second week of May and if you book now, Mariotey Hotel is offering a 25% discount on rack rates. Should we look up some flights?
Imagine the tremendous power that was just unleashed. All this makes travel planning a coherent and pleasant experience. This is the direction where Bots need to head. Instead of being a rule driven engine, they need to be able to process facts and come up with recommendations and solutions. Easier said than done, but with Machine Learning and Natural Language Processing, this is the way forward
Ecommerce:Amazon started selling books when it started. The rest is history. It has become the de-facto choice of online shopping. Their investment in technology and solid foundation is second to none. What chance do Bots have to disrupt this?
The answer is that Bots don’t have to disrupt Amazon or any e-tailer in particular. Amazon sells a lot of things; it doesn’t sell everything. It delivers quickly with Prime but you may still need something quickly from a local store. You may need a price check across e-stores and then decide, but all of this takes time. That is one thing the availability of which varies wildly from one individual to another but is generally scarce. Imagine the ability to say: Order me the supplies for Anna’s birthday this weekend. The obedient Bot goes about compiling a list of items that need to be ordered (from previous event or a preferred list), compares prices from vendors, places order and gets the ducks lined up for deliveries and pickups. Now, that’s disruption. It doesn’t matter where you order from – it’s supreme convenience that’s being added and mundane repetitiveness going on the chopping blocks.
Of course, there are scenarios that need to be take care of and edge cases that may present themselves with such a complex system, but it’s never been a hindrance for innovation. If cars could drive themselves, a cake could be ordered by a Bot without having to worry about it.
Retail: Retail has been a traditional Sales oriented gig. You need to staff counters, lay the material on the shelves, dress mannequins, and maintain elaborate inventory management solutions. Is there an efficient method that removes some of the complexity from this or have we hit the edges of our imagination?
What does it take to create a compelling retail Customer Experience? Maybe, use Beacons to track Customers, their preferences, have the things ready, pre-emptive preparation of monthly groceries, track the frequency of wardrobe usage, dresses worn most or the colors that are bought most. Will all of this happen in silos? Or, is it one connected experience?
Bots could direct the consumers to the right aisle based on their physical characteristics to avoid being surrounded by incorrect fit apparel or shoes. They could keep a tab on promotions and other sales events and get you the best bang for the buck by directing your attention to the most attractively priced items.
The disruption is around the corner. Retail as we know it is going to change dramatically. Fuel prices change multiple times a day, what’s to say it can’t happen with retail prices?
Banking: Money is the currency that controls our commercial existence. It replaced barter system and is the language that transcends geography. Quite a few of our daily actions and decisions are influenced by money. Banks are the single biggest stores of money. They are guardians of our wealth. Since they are the only agency authorized by the society at large to control that, it is natural that our interaction with banks is serious in nature and happens frequently. Bots could play a key role here. Banks have already started to deploy Bots to get the routine affairs done and dusted without human intervention. Mobile apps already gave you access to your account information, transfer funds and place requests. Bots take it a step further by letting you open bank accounts, get loan offers, apply for credit cards, get the best rates, book deposits and more.
Bank of America recently launched its AI Bot, Erica. It dispenses financial advice on how an individual could save money, investing in something, changing the credit card, paying off balances etc. The Bot uses artificial intelligence, predictive analytics and cognitive messaging to help customers do things like make payments, check balances, save money and pay down debt. The objective is to extend some of the benefits of the one-to-one personal service and advice usually reserved for top-tier customers. It remains to be seen how far this would go but appears to be a good first step. The trick would be to deliver accurate financial advice – no one likes to lose money following a bad tip!
Q: What design aspects should be included in ‘bot frameworks for businesses’ that take good care of custom requirements of a particular company?
A: The simplest way to gauge this to avoid the pitfalls that exist in the traditional modes of communication – e-mail and voice. Bots need to remove all the artificial barriers that get put into place because of how we know things to function. All assumptions need to be looked at with a fresh outlook and then design the experience around the customer. Bots are one of the many digital touch points available to the customers. They have to be the cynosure of those touch points. Some of the points to consider are:
Contextual awareness: If the customer has had e-mail or call with the business support team earlier and is reaching out again through a bot, the bot needs to be aware of this and act accordingly.
Non-transactional relationship: A bot conversation cannot be treated in silo. This compliments contextual awareness and add on to it by merging various facets of the business – sales, support and recommendations. The customer should not be expected to re/start or abandon a conversation because they are dealing with bot running in “support” mode rather than “sale” mode.
Universal interface: The customer needs to be able to complete their activities through a unified interface. They shouldn’t be redirected to a browser to pay or the bank site to verify funds. This is critical for commerce oriented bots.
Relevant recommendations: If you’re selling wine, it makes sense to engage the customer is cheese offerings as well. There would likely be validations for age and delivery areas as well – all those need to be accounted for and in a painless, frictionless UI.
Customer Experience: All of the above needs to tie into a seamless Customer Experience. If the customer is not at the center of the journey that has been designed, the churn rate would be high.
Q: What is “not a chatbot’s cup of tea”?
A: How about – they can’t read our minds. Right? Of course, they can’t. The success of any new technology is contingent on right use cases and adoption. If we taper our expectations in terms of what Bots can do and focus on the right problem statements, then yes, Bots would fit right in. Most of the underlying technology powering Bots is in infancy and needs time to reach commercial maturity. The technical talent required to harness that technology is scarce as well.
Majority of modern use cases of Bots or the ones already out there are nothing more than if-else rule conditions that follow a pre-defined flow path. Each question to the Bot is a decision box, you either fork left or right. You could end up looping infinitely with the Bot since there are only so many rules that you can bake into a product. The trick is to teach the Bot to learn, understand user behavior and have contextual awareness. If you’re chatting with a Bot, leave the conversation and circle back the next day – the Bot needs to know what was last discussed and take the conversation from there. This is something that is currently not happening and plays a vital role in creating an engaging Customer Experience. Bots also cannot take policy decisions reserved for humans, which probably require sign-offs and following a process.
It’s not all bad news though. Understanding the limitations is an important step in unleashing the true potential. Once we know what not to expect, we’re able to focus our energy on things that will actually yield meaningful outcomes. Applying Machine Learning, Analytics and NLP (Natural Language Processing) to the conversation would go a long way in providing a rich experience to the interaction. It’s not easy though. However, as this pivot happens, the relevance of Bots increases manifold and they continue to become an integral component in a company’s digital touch points stack.
Q: When is a good time to adopt a Chatbot in your business ecosystem?
A: There is always to a time to market and opportunity cost. Bot is one of the digital touch points. If a business is content with a brick and mortar store, only wants to accept cash, shuns the concepts of electronic commerce and does not believe in placing the customer in the center of their ecosystem, then they would never need a bot. If any of the above are false for a business, then they need to take a serious look at bots.
Bots place the control of the communication firmly in the hand of the customer and once a business starts designing journey with customer in mind, then Bots make a natural fit. Yes, the technology will take a little while to mature but if you wait too long, then you lose the opportunity. Getting too late to the party will make you “also ran”. Businesses that were early adopters could easily show their prowess with a scripted bot. The bar has already risen higher. You need more than that now. You need a bot that can provide context and connect the dots. Innovations such as Apple’ Siri and Amazon’s Alexa have raised the expectations that people have from conversational interfaces. There is no silver bullet that you will take a business to the pinnacle of quality and customer loyalty. Bots can aid business in their quest for customer retention, enrichment and advocacy.
Businesses usually offer a transactional relationship, especially in e-commerce. Chatbots make the tedious task of completing the purchase process easier by addressing the customer’s queries that would otherwise have required a human to be inserted into the equation.
What draws businesses to chatbots is the fact that they open new avenues for the business to interact with their customers without trying to upsell products, losing temper or being unavailable. Once Natural Language Processing (NLP) and Speech Recognition goes mainstream, bots would manifest themselves in a different form and take pre-emptive actions that would preclude a customer from taking repetitive actions.
If you are considering ChatBots as part of your Digital Strategy to improve customer or employee engagement, this is an area where we should be able to help.
You can find more on how to leverage chatbots for your business by reading our latest brochure on the rise of chatbots.
Schedule a call with our Chatbot expert to get more clarity on the Chatbot strategy and implementation. Write to us at firstname.lastname@example.org.