By Mark Jancola, CTO & VP of
Engineering at Conversica
(AI) is transforming all sectors of the economy. While AI will radically alter
how work gets done and who does it, the technology’s larger impact will be in
complementing and augmenting human capabilities. Many companies have already
leveraged AI to automate processes, but the next frontier is applying AI
technologies like Conversational AI solutions to directly interact with
prospects and customers and deliver a better customer engagement experience.
The companies that use AI in this way are the ones that will build the next
generation of intelligent businesses – where humans and Conversational AI
solutions work side by side as a team to divide up the work. As a result,
customer-facing teams can build trusted relationships and solve complex
problems while the AI solutions automate the repetitive, routine interaction
that bog people down.
While most AI is not a
plug-and-play technology, even those with quick ramp-ups that deliver
immediate automation results need ongoing care. After that initial burst,
enterprises must continually teach these solutions to improve performance in
order to become strategic assets. AI needs constant “interaction” with human
counterparts. Consider Conversational AI technologies, such as virtual
assistants, digital agents, and ‘bots’, technologies that enable human-like,
two-way conversations with customers across various communication channels.
These solutions require continuous learning and direct input from the
organization on nuances in structure and data to deliver the desired results.
In fact, if applied and nurtured correctly, Conversational AI solutions have
the potential to extend far beyond today’s use cases.
Now you’re speaking my language
A key element of
Conversational AI technology is Natural Language Processing (NLP), which
permits a software’s ability to attach meaning and action to human language.
Each organization has its own unique vernacular that Conversational AI
solutions are required to learn. For instance, some companies may call
employees associates, while others refer to them as partners. In medical
organizations, customers are called patients, and clinicians are providers.
Similarly, products could have distinct names or features known and understood
by their customers but may not be commonly recognized on a wide-scale.
In order to facilitate
seamless conversations and be trusted by customers, Conversational AI solutions
need to be configured to speak like a member of a specific organization.
Collecting data and building a taxonomy is necessary for organizations to
successfully teach Conversational AI to handle idiomatic expressions or make
recommendations in a natural way.
It’s not a one-and-done process. Training on
company-specific language requires an inventory and a systematic method to
curate terms. Conversational AI solutions are continually evolving as
enterprises adopt new phrases or transition away from others. Like any new
employee, these AI-powered team members should receive feedback from humans to
improve their effectiveness and accuracy over time. It’s essential for
companies to learn with
Conversational AI solutions, as trends in emerging markets may lead to new
consumer behaviors and process improvements that bear new knowledge. This
information needs to be fed into both the human workforce and the AI system to
react appropriately and become more accurate in responses to newer questions and
utilize Conversational AI technology for data capture and analysis,
simultaneously informing humans and machines to improve future actions as part
of the training loop. The human language evolves quickly; consider the phrases we
use today that were not part of our lexicon a year ago — Zoom, PPE, social
distancing, meme stocks. These phrases work their way into language and need to
be captured, recognized as a trend, and configured into the Conversational AI
system to become part of the current communication paradigm. How would a
customer react if they requested to schedule a meeting over Zoom and a
Conversational AI solution responds, “What’s a Zoom?”
Meet people where they are
interact with companies over numerous platforms. An interaction with a customer
and a Conversational AI solution could start on a website and transition to an
email exchange that leads to a phone call with a human staff member. When over 75% of shoppers use multiple
channels to make a purchase, Conversational AI solutions and workforces as a
whole must be able to switch between various modes of conversation to ensure a
seamless customer experience. Nothing is more frustrating than providing
information on one channel and then having to repeat it on another. A
customer’s digital experience should be designed as if they were communicating
with a human, with each interaction resembling a 1:1 conversation that is
context-aware and informed by past interactions. Understanding a customer’s
communication style and suggesting the best next step is critical for
businesses to successfully deliver an impactful customer experience.
Good enough is not good enough
Customer trust is every
organization’s utmost priority, and enterprises should hold Conversational AI
systems to the same, if not higher, standards as humans. A quality focus means
that there have to be thresholds and guardrails in place to elevate
conversations from Conversational AI solutions to humans. One nonsensical
response from a Conversational AI solution can undo trust in the service
process and have a direct and lasting impact on customer experience.
As a customer feels a personalized and human
connection with a Conversational AI solution, it becomes even more
critical that the machine get responses “right.” Furthermore, Conversational AI
solutions need to be trained to have the same respect and sensitivity for
privacy as human associates do. Conversational AI technology must be configured
to handle uncertain situations gracefully and should have the ability to
quickly bring a human into the loop to confirm a response or to take over a
Making an impact on an
enterprise’s bottom line requires more than just automating manual tasks.
Achieving a business outcome is more complex and involves automation solutions
to combine humans’ critical thinking and machines’ sheer computing power. As a
result, systems need constant attention to understand the business’ language,
stay relevant, and meet users’ omnichannel needs. Much like a parent,
enterprises that understand their role in nurturing the Conversational AI
solutions will create a new digital team member out of the 0s and 1s of a
Mark Jancola currently serves as the CTO
and VP of Engineering at Conversica, Inc. Jancola previously served as Executive VP Engineering & Operations
at Apptio (a Vista Equity Partners portfolio company). In this capacity,
Jancola established worldwide product development operations across multiple
sites in the US and India. Formerly, Jancola served as the GM of the Apptio Digital Fuel business
where he was responsible for all facets of business operations (P&L responsibility)
and acquisition integration.