Marketing 0

Human and Machine in Customer Interface
In the customer interface, humans and machines can go hand in
hand, too. Typically, channel selection depends on customer tiering.
Interactions with humans are generally reserved for hot prospects
and most valuable customers due to the high cost-to-serve.
Meanwhile, machines are employed to qualify leads as well as
interact with low-cost-to-serve customers. The service segmentation
allows businesses to control costs while at the same time managing
risks.
Indeed, the use of AI for interactive purposes is risky. Microsoft's
now-defunct chatbot Tay illustrates this. Learning and responding
from abusive tweets from provocative users, Tay began posting
equally offensive messages via Twitter. The bot was retired after only
16 hours since launch. Google experienced a similar problem when

---

its image recognition algorithm labeled a user's black friends as
gorillas. The company fixed the algorithm by eliminating the word
gorillas from its labels altogether. The insensitivity of AI is one of the
biggest threats to manage.
Computers are a good fit only for predictable inquiries and
programmable tasks. Solutions such as self-service kiosks and
chatbots handle only basic transactions and queries. People are more
flexible across a broader range of topics, and therefore better suited
to perform in a consultative role. Their superior contextual
understanding allows humans to adapt to unpredictable
circumstances and unusual customer scenarios, beyond standard
procedures.
Software company HubSpot, for instance, uses a chatbot to capture
and nurture leads in its top and middle sales funnel. But the
company assigns a sales force to do consultative selling for qualified
leads and a high-touch team for onboarding. For post-sales, the
company returns to the chatbot to answer simple queries.
Above all, humans are warm and friendly. For any tasks that require
empathy, human-to-human connection offers the best solutions.
Even some companies that have already installed high-tech customer
management solutions still rely on people's social skills for service
deliveries. Take Marriott with M Live, its social listening center.
When social listening identifies a missed opportunity in one of
Marriott's properties—for example, a couple on honeymoon—the
command center notifies the respective hotel so that they can
surprise the guests.
Understanding what automation and a human touch can deliver is
the first important step to designing an excellent Omni CX (see
Figure 7.3
). And it is often not about selecting one or the other.
Businesses need to drop the “machines replacing humans” mindset
or risk missing an opportunity to optimize their operations. In truth,
humans and computers should coexist and build on each other's
strengths in most touchpoints. Thus, the next step involves
reimagining and redesigning the customer path to harness the full
power of the collaboration (see 
Chapter 11
).

---