AI research in terms of architecture / implementation has so far primarily been data-based endeavors, with relevant and effective collection of data and their organization considered the chief building blocks for an effective AI-model. With both customer-support as well as customer-experience being at the center of any customer-driven business enterprise, the future belongs to the construction of effective customer-centric models. The future is also about devising models that not only smoothen the process of interaction (between the customer and the business in question) but, given the altruistic nature of innovation in the sense that technology is easily replicated and there's a short shelf life attached to any first-time mover, businesses will need to quickly figure out at the same time a competitive edge for themselves in order to stay relevant. To my mind, the enterprise that shall break the clutter and carve for itself a niche is the one that is able to integrate 'emotion' with 'effectiveness.' And how to get there?
Part 1 – Building a Directory of Dissatisfaction
Let's take, for example, any contemporary business that has shifted from human voice-driven protocols to AI communication tools, like 'assistants,' or 'chatbots,' to serve as a platform between the said business and its customers. While a cold, calculated response system is more geared towards shuffling out pre-determined 'lines of communication' akin to their human equivalent that previously referred to a handbook categorized in an 'if-then' manner, it is the enterprise's ability to handle unlisted 'unique situations' that does and always shall make a difference.
Say a perishable item purchased from an e-retailer recently shifted to an AI-driven interface receives a complaint not listed in the AI's data bank. Previously, in a fairly efficient organization, a junior human associate handling the company's back-end ops could be impressed upon with the personal nature of the problem, and a good organization would empower the said associate to immediately contact his empowered superior to send a replacement and / or provide an on-the-spot money-back offer with an added bonus to keep the said item. The decision is taken at the spur of the moment, maybe using photographic evidence of the damaged item, judging from the customer's tone of voice (measuring 'sincerity'), his past purchase record, and taking a long-term view of the customer's intentions and the company's willingness to invest in him / her. It may as well be the company's strategic goal to treat such complaints / issues with a view to free inventory space and treat such a loss as an incentive to retain customers. In this case, the customer doesn't just get his money back, but he even gets to keep the damaged article, while the monetary nature of loss notwithstanding, the business, as stated previously, gets to retain a satisfied customer, thereby transforming a negative experience instantly into a positive outcome. The value of the damaged item and the monetary loss notwithstanding, the company is well compensated in terms of brand recall and free word of mouth travel, considering a good customer will refer to his positive experience and keep coming back knowing he will get a fair transaction.
Every complaint or query is therefore an opportunity to enhance our understanding of the customer experience. It also helps us adjust the company's delivery systems accordingly, provided our response systems are not confined within a 'set list response framework.' We leave nothing to chance, and we leave no stone unturned to retain a customer dissatisfied with a negative impression, which, over time, instead of denting the brand image, turns into brand capital. Trouble starts when the advent of AI-tools makes it imperative for an enterprise to shift its back-end ops to auto-voice-generated platforms.
While a data-based set-tool based on an 'if-then handbook' can undertake most queries / complaints, what replaces the previously empowered junior associates / or his seniors' on-the-spot decision? How to empower and educate an AI tool to handle such unique situations not part of the data set it is made of? In other words, how to transform an unempowered AI tool into an empowered associate and take on-the-spot decisions considering the said situation in not mentioned in the 'if-then handbook?'
The key to my mind is for any business enterprise to undertake a long-term phase out than a knee-jerk replacement of its back-end ops with AI tools. Aim should be to—
- Keep a minimum of 1 year (or its multiples, as the situation or complexity of the business demands) between the goal and its implementation
- Use the time-gap to create a Directory of Dissatisfaction, which is different from the traditional satisfaction-index. Meaning, while the satisfaction-index is more of a customer-entered survey data, it leaves out major interactive indexes such as the 'uniqueness' of the complaint and the way the customer wants the problem to be resolved (rather than the company wants it to be done). As such, the unstructured data in terms of the voice call, and subsequent interaction between the customer and the associate, and the way the latter 'solves' the formers issue may be transcribed as a data set and its voice data archived, allowing for the developers to extend the role of the ai-tool to pick, absorb and assimilate the associate's response system.
- The more such'unique' cases, the more the said AI tool picks up its 'emotional intelligence,' giving its developers ample time and opportunity to bug-check, trim, and size the said tool.
At the same time, I also propose a 'natural evolution of AI tools' similar to training a junior associate by placing him in a real-life scenario with a hovering supervisor around him who jumps in at each escalated juncture and solves the customers' problems, giving the associate time to learn. Once the company has outlined for itself an adequate measure of the Directory (of Dissatisfaction), the second stage of the Time-gape may be taken by—
- Gradual introduction of the said AI tool into the interface in place of the human associate.
- A team of human associates (depending on the size of the 'interactive transactions per day' per business maybe picked from a pool depending on their personal efficiency (measured effectively by their problem-solving intellect) and promoted to 'Interactive Supervisors' and their original supervisors, the most efficient amongst them given a jump up the ladder as 'Business Supervisors'] whose chief job remains to serve as mentors to the team of Interactive Supervisors as well as to implement a 'similar evolutionary transformation' in other sectors / regions of the business.
The said business may therefore be a model implementor of AI-transformation by retaining—
- The best personnel from its human resources in terms of their problem-solving abilities and providing them with a measured growth chart, turning them from 'Implementors' to 'Ideators.'
- Customers that are not left to deal with'plastic-responses' but owing to a more 'personalized response systems' they become part of the enterprise's long-term 'Inventory of Satisfied Customers.'
Part 2 – Mapping of an Inventory of Satisfied Customers
The customer is king. He serves as the most precious capital of every customer-centric business enterprise. Rendering of services and the resulting imprint of experience in the customer's mind is what determines the fate of every service-sector business. The monetary gain emanating from the introduction of AI-tools to replace service roles previously handled by their human equivalents runs the risk of being inversely proportional to the satisfaction index of a company's customer base if a said transaction is accompanied by a mechanical relationship management owing to a hurriedly integrated AI-interface. Any non-human interaction is prone to appear plastic, staid, or cold if it is unable to manifest into that one thing that bonds a customer to an enterprise—'RELATIONSHIP.'
Warmth and fairness are chief elements in a business transaction. The way the transaction takes place, the satisfaction of its intention, and a parity between what the receiver expects and what the giver can provide is what constitutes satisfaction in which services are rendered in return for services asked. This is where the human part plays a key role in every service sector. The ability of a business enterprise to show its human face, place in front of its operations a human presence that the customer, who is in effect a human presence occupying the opposite spectrum, marks the culmination of a transaction or an enquiry.
We, as humans, are born to trust humans as well as the presence of everything human. To find such interaction or interface suddenly turn mechanical is one of the formidable challenges felt (or about to be felt) by every business enterprise in line to introduce AI-tools in a service-oriented business. Once again, my prescription for such businesses would be to—
- Have a reasonable time-gap between the goal and its implementation (determined by experienced human interlocutors).
- Use this time-gap for creating, along with a Directory of Dissatisfaction (dealt with previously), a process of measuring our customers, in other words: scale customers according to their worth. While a regular customer with substantial interactions with a business may demand personalized attention vis-à-vis a periodic customer, a first-time customer with moderate purchase power (or of moderate worth) is equally important to retain, and it's this foresight that must have the business provide him with a service that has him coming to the business again and again.
- As such, the next step would be to scale customers as per their potentiality to develop and come up with strategies to ensure their moderate worth is transformed to high worth.
- In the course of time, develop efficient psychographic profiles of each customer, mapping their likes and dislikes.
Once an inventory of a well-mapped customer base is complete, the next time-gap maybe utilized to vitalize the mapping by fine-tuning its findings, checking, and double-checking every trait and habit using a fine fusion of AI tools and human inputs.
The end product of such an exercise would be to ensure that the business has, at the end of the time-gap a formidable list of a varied customer base. Once that part has been accomplished, it is time to introduce specific AI tools to ensure that each high-value customer is provided with optimum service, but without totally cutting away the human interactive platform. Meaning, managers and associates must continue interacting with them as before, without replacing the interactive platform by ai-tools. Rather, they should be furnished with an auxiliary AI experience (read: supporting) to enhance the transactional experience by using respective custom-made programs to identify their ever-evolving needs and requirements, as well as find ways to satisfy them. Customized programs may, in turn, be used for new or frequently interacting customers to optimize their experience and ensure they are made to feel their worth, so they are approached with warmth.
In other words, AI tools may prove to be a significant value addition to new or existing businesses dependent on human interactive platforms, but never at the cost of totally eliminating the human-part from the interface. Each can, however, prove effective in multiplying the other's capabilities and worth and help a business to retain not just its customers but also to choose and calibrate its most efficient manpower from the overall HR-pool for a balanced and fruitful future.
To my mind, the 'either / or scenario' considered by business enterprises in the face of new AI-technologies and its human-pool serves as a weaker option when compared to its more efficient 'human / ai fusion enterprise.' After the initial euphoria and the rainbow of AI settles down, we shall find that both human and its ai-counterpart works best not in isolation, but as an entity that complements one another.
