Automated technologies have advanced tremendously in the past decade, sometimes so quickly that we’re caught off guard by their accomplishments. For many, the question is often, “How far can automation go?”
Mohanbir Sawhney, a professor at the Kellogg School of Management, recently explored this topic from a marketing perspective. He visualizes a future where brands use big data to categorize every possible customer interaction and use AI algorithms to generate specific responses in real-time. Sawhney refers to this experience as sentient marketing, and believes it’s closer than we think.
What is sentient marketing?
Sentient marketing is Sawhney’s term for any set of processes that enables enterprises to deploy “personalized customer engagement at scale and in real-time.” It is a practice of customer interaction rooted in automated systems, similar to how algorithmic trading is conducted on the stock market.
Despite the Hollywood imagery, the phrase conjures — no, sentient marketing doesn’t have to be self-aware — it’s really not so different from the way many traditional campaigns operate. Today’s marketers use automated processes to send emails, forward customer service calls, populate events calendars, and more. The distinction is that sentient marketing can go further than automated tasks typically carried out by the average brand. Once a process is formalized into a series of data-based objectives, a sentient marketing service could largely run itself.
What does sentient marketing look like in practice?
From Sawhney’s perspective, we can streamline sentient marketing into today’s technologies. Each element of sentient marketing — from chat and email bots to big data algorithms — already exists in dispersed forms across various organizations. The only reason we haven’t amalgamated these processes is because marketing institutions aren’t willing to bring them together.
One of Sawhney’s primary examples of what sentient marketing might look like comes from Reliance Jio Infocomm, a digital services company operating in India. The organization currently serves 250 million customers, which means it’s not unusual for its various contact centers to receive 600,000 to 800,000 customer requests per day — many of which are requesting contract upgrades. Processing these requests using human operators is an immense task, but Sawhney’s sentient marketing system could respond independently to most calls.
In fact, Jio is working on such a system, referred to as a “real-time promotional response system.” This process starts by automatically scanning for contract requests through all marketing channels before determining which ones are legitimate, relevant, and valid. From there, it runs the request through an algorithm and decides whether it makes business sense to negotiate with this customer or let them go to competitors. Finally, it either accepts the request or assembles a counter-offer based on some kind of promotional pricing option or additional service. If done correctly, interactions occur in real-time without the need for human confirmations.
Sawhney isn’t alone in preparing for this future. While the phrase “sentient marketing” is his brainchild, everyone from marketers to tech giants is experimenting with automated processes. The immense volumes of interactions brands must respond to each day requires some level of data management to support and engage customers. Automated processes that respond in real-time would offer any organization a strong advantage over the competition.
Professor Mohanbir Sawhney, Clinical Professor of Technology teaches a class at the Kellogg School of Management in Evanston, IL. Photo by Eddie Quinones.
What are the drawbacks of sentient marketing?
In his article, Sawhney doesn’t mention any potential disadvantages for sentient marketing. His argument largely states that because big data can improve customer engagements, it will improve customer engagements. In his view, it’s simply a matter of capitalizing and expanding upon existing infrastructures to lay a groundwork for automated real-time marketing efforts.
But while brands certainly are still catching up to automation’s potential, the technology requires significant refinement to operate at scale. Consider the following real-world examples of automation challenges faced from tech-forward industries:
- Chatbots used to automate recruitment processes lack emotional context, struggle with language barriers, and can be targeted by malicious actors.
- An algorithmic trading program in Hong Kong breached the stock market’s code of conduct, leading to $4.5 million in fines.
- A Chinese facial recognition program flagged a billionaire’s face on a bus advertisement as a jaywalker.
While any new technology faces challenges that eventually improve over time, no automated process is ever error-free. And just as the consequences of mistakes become amplified during highly complex tasks like advanced algorithmic trading, brands will have little room for mistakes when interacting with 800,000 customers in real-time.
Taking a closer look at Sawhney’s Jio example, it’s easy to imagine a scenario where a coding error gives some customers a 99.99% discount each day. If Jio is interacting with almost a million customers per day but the sample size is too small to be spotted by human operators, this problem could compound quickly in a short period of time. After a certain point, it would impact bottom line at worst, or be a major embarrassment at best.
What can we learn from sentient marketing?
As a practical reality, we probably won’t see sentient marketing in the near future. Yet as a concept, there are certain merits and principles to consider. The automated technologies powering Sawhney’s vision of sentient marketing can already reduce inefficiencies, engage with customers at scale, and provide highly personalized interactions. The focus should be on determining how automation is most effectively and safely deployed, instead of fully embracing it on principle.
It’s likely that one day, sentient marketing as imagined by Sawhney will arrive. But whatever happens, marketing doesn’t need to be sentient to be revolutionary. Whether campaigns are automated, sentient, or powered by human operators, finding thoughtful ways to better engage with customers should always be the primary goal.