Chatbots have become an integral part of Customer Relationship Management (CRM), especially, when it comes to handling customer grievances. In past (before 2014-15 era), bots were used guiding customers to the customer care executive, who would listen to them and respond accordingly.
Nowadays, a machine-learned response generates automatically whenever a consumer sends a digital complaint. But with a machine trying to analyse the situation and respond, often, the algorithms go wrong.
Hence, starts the endless loop of increasing customer frustration and useless templated replies that go on to show the inefficiency of teams dealing with machine learning in the customer support across the ecosystem.
Customer Care pages of large tech startups on Twitter are majorly run by chatbots, and the results are hysterical, be it Zomato, Ola, Flipkart, BigBasket or Paytm.
Entrackr has studied a recent support conversation on these pages and tried understanding common issues faced by the customers due to the automated replies.
The basic issues on payment portals that trouble users are transaction failure amount getting deducted but not adding to the wallet, unsuccessful bill payment, etc.
With the exploding increase in the number of UPI-enabled payment platforms and their users in the past few years, the competition has escalated. To perform better the portals engage in practices like an introduction of offers and cashback. Unfortunately, customer services have also become worse.
Paytm has the highest MAU in the payment gateway space and the worst customer care service. Customers are openly expressing their frustration with the scripted replies.
Going through endless conversations with the machine, they either receive template resolutions or none at all.
Sometimes the bot either asks for the conversation to move to the direct message (DMs) or keeps repeating the detected fault instead of providing the solution.
On Google’s platform Tez, the replying machine sometimes doesn’t even respond to customers in plight. For Instance, a person paid Rs 10,000 as his electricity bill via the interface but his payment didn’t reach the department even after the transaction took place. He went through the customer support chat, raised a dispute and when he got no response for 9 days, he took twitter’s help. And even there, he received no reply.
“Thanks for sharing the same. We have raised this to the concerned team to get it sorted on priority”, is the standard statement by the Mobikwik automatic reply machine. That, or asking for details via DM without a guarantee of response.
There is also an instance where the bot basically left it upon the customer to solve his problem by connecting with a third party.
The company also receives heat for canceling accounts without reason or notice.
PhonePe’s bot either tells the customer to wait and then thanks them for their patience or asks for a DM.
Government undertaking, BHIM, is infamous for not responding to customers in peril. Countless tweets remain unanswered. The few tweets that are considered, the users are asked to switch to DM, and then they go unresponsive again.
People on Flipkart and Amazon are fed up by the delayed deliveries, missing items, wrong orders, delivery person’s behaviour amongst other things. But to top it all, the customer service isn’t helpful either.
An Amazon customer’s chances of getting its problems resolved via Twitter isn’t high. The standard most used statement is that the “relevant internal team is looking into the matter.”
On Flipkart’s support page, the bot, repeats the same response for everything, and it turns out to be highly irrelevant at times leading to no resolution.
Online groceries such as BigBasket and Grofers face a particular challenge of delivering the correct item within a short duration and upon failing to do so, they receive a backlash by customers.
BigBasket’s twitter ARM just reassures the customers by telling them the team is looking into the matter, whilst the party in trouble often receives no relief.
Grofers’ chatbot is more humane but equally helpless, as it doles out a very heartfelt apology with lots of serious adjectives after asking for DM, but in the end, customer’s problem isn’t solved.
Foodtech majors Zomato and Swiggy also lack expertise while leveraging artificial intelligence in handling consumers’ concerns.
For instance, a Zomato customer got perplexed when he did not receive promised rewards, and tried to get help via chat support, but instead got blocked. He then resorted to Twitter and received ambiguous replies. His account was unblocked but, the original question remained unanswered.
Swiggy consumers face immense disappointments when their orders are cancelled due to technical glitches, without intimation, when all they have done is a complaint against the delay in delivery.
Even the restaurant owner had to spend 3 days wrestling with the bot to get a quick resolution, the final result is still unknown.
However, it is worth mentioning that Swiggy’s responses are still more relevant, and often do help the consumers in resolving issues.
Ride-hailing apps – Ola and Uber have to deal with customers facing issues like irrelevant charges, wrong routes, drivers’ misbehaviour very often.
Relatively, Ola’s support mechanism is slightly better than many, with its quick responses that are circumstantial and relevant, and the practice of clarifying the issue and providing the resolution publically after the vulnerable information is shared over DM.
But on the other hand, Uber provides an unsatisfactory machine operated support channel to its customers. Here also the consumers are stuck in the cycle of DM, and their matters getting looked into by the relevant teams, and other irrelevant responses.
Amongst the consumer-facing services, telcos are worst as far as solving queries of customers is concerned. Be it Airtel, Jio, Idea or Vodafone, all companies face a barrage of complaints on Twitter by customers over recharge failures, cash-backs, network, and other issues.
Very often, these telcos look hapless while replying to furious and helpless customers via bot.
Conclusion: Bots have to go a long way
While the machine learning, AI, and chatbots have made the entire process of customer support much easier for the companies, consumer-facing Indian Internet companies are yet to cover a long distance in solving customers’ concerns through chatbots.
Since past 18-14 months, there has been a lot of buzzes around AI and every technology-enabled company claims to be practicing it, the scope of improvement and optimisation in dealing consumers’ concerns is huge as well as the need of the hour.
With the improvement in executing artificial intelligence for customer support, companies can definitely assist customers with more agility and to the point. Going forward, it would be interesting to see how aforementioned companies optimise machine learning to address consumers’ real pain and concerns.