Technology is cool, but the casual things that humans do are incredibly difficult even for tech to replicate !

Nimisha S, Conversational AI Designer at Deloitte, specialises in Gen AI, NLP and Conversational AI technologies that enable chatbots help customers accomplish tasks.

Nimisha talks to  Shyam Krishnamurthy from The Interview Portal about challenging and dynamic problems that she faces on a daily basis while designing chatbots that can comprehend human language.

For students, in the age of AI, strong language skills give you the edge that can help in training bots to understand the nuances of human communication.

Nimisha, can you explain your background to our young readers?

I grew up in Calicut, Kerala. I was a Science major for my 11th and 12th with IP (Informatic Practices) as my optional subject. I learned the basics of coding and IT from there. I was very active in school events. Later, I changed my plans of pursuing Engineering to English Literature. My parents are from ordinary backgrounds, with no connection to IT industry. I never thought about career back in school, I just enjoyed whatever came to me, did well in school events, and took up leadership roles. I chose courses based on my interest rather than the thought “will I get a job if I take this?”. Though it worked well for me, how you choose your career depends on a lot of individual factors based on your economic class. Generalizations in this matter is not advised. ☺

What did you do for graduation/post graduation?

I graduated from Farook College, Calicut University and finished my Post Graduation from The University of Delhi both in English Language and Literature. I have also qualified NET, English and am eligible to be an Assistant Professor of English Language and Literature in any colleges/universities in India. 

What were some of the key influences that led you to such an offbeat, unconventional, and uncommon career in Conversational AI?

I accidentally ended up in this career. I had put out resumes on Job Portals like Naukri and Indeed.co.in back in 2020 soon after my post-graduation. The intent was to do something and gain outside experience. Content writing, my expertise in language and voice overs (a part time freelance work I did) were the keywords on my resume. A CAI company hired me to record the voice of their Automated Voice Bots in my regional language, Malayalam. I started with voice recordings for AI powered bots, which eventually led to editing the bot scripts (how the bot talked to users) and training the bot with human way of speaking. Once I got into the field, I explored different areas of work within this field and stuck to the area that interested me the most and that paid the most, Conversation Design. 

CAI companies (Conversational AI companies) are companies that create customer service chatbots and voice bots. Instead of humans, it would be automated bots that help users through multiple requests.

These bots are human support agents which help customers with various actions across domains. For example, in banking, a bot can help a customer transfer money to another account. Or a bot in a hospital can help a new patient book an appointment.

How did you plan the steps to get into the career you wanted? Tell us about your career path

As I explained above, this was an accidental discovery, and since then, it has been a major experience in Conversational AI  that defined my career. 

At Gnani.AI, I started as a Linguist. To explain my role, lets take an example. Lets say there is a bot helping the user pay EMI and the user is telling the bot that he will pay it next week. The bot needs to be trained to understand human language because its a software. We enable the bot to understand human language by training the bot with lots of data and variations of the same messaage because humans can convey the same message in different ways.

As Linguists, we create many datasets of different ways of saying the same message and have to train the bots on root words, subject, verb, adjectives , synonyms so the bot understands when a customer utters a message.

Next, as a Test Analyst, I would talk to the bot just like a customer would, and test the bot’s response to different variations of the same message. The idea is to ensure that the bot behaves in expected ways by either acknowledging the information from the customer or asking follow up questions if needed. This helps iteratively train the bot to understand messages from the customer.

As Conversation Designer, I design the user journey. What I mean by user journey is, defining a particular activity and a conversation that happens between a customer and a bot. Lets say a customer wants to return a order. How would the conversation happen? The customer types in a message and the bot responds by asking for the order Id. Once the customer enters the order Id, the bot confirms the order with details and once the customer acknowledges it, the bot initiates the return process. This is an example of a Conversation Design based on the kind of user request.

A Conversation Designer also defines the function of a bot depending on the domain . For example, if it is an e-Commerce or shopping website, what would be the reasons for a customer wanting to chat. It could be to enquire about an order, place an order or return an order or ask for a refund. I would think about the reasons for using the bot and define the purpose of the bot.

I was again a Conversation Designer at my next job at Kore.AI. Kore.ai is a company that develops enterprise-focused conversational AI and GenAI products.

Along with that, I also attended online Conversation Design courses. I became active on LinkedIn and followed Conversation Design communities from all over the globe. This helps me to stay up to date in my field. I also started openly writing about my learning experiences and challenges on LinkedIn which gave me access to others who have been going through similar journey through which I understood how different companies work. 

How did you get your first break?

A recruiter called me after finding me on Naukri for voice recording.

I had put out resumes on Job Portals like Naukri back in 2020 soon after my post-graduation. The intent was to do something and gain outside experience. Content writing, my expertise in language and voice overs (a part time freelance work I did) were the keywords on my resume. A CAI company hired me to record the voice of their Automated Voice Bots in my regional language, Malayalam. I started with voice recordings for AI powered bots at Gnani.AI, which eventually led to editing the bot scripts (how the bot talked to users) and training the bot with human way of speaking.

What were some of the challenges you faced? How did you address them?

Challenge 1: 

Understanding the terms associated with the profession. I asked around, researched on the internet and reached out for help from colleagues

Challenge 2: 

Design challenges . Problem solving through design thinking was a new skill. Since I never entered this area through designated courses, I had to learn everything on the go, by making mistakes. Every learning is experience oriented for me. 

Challenge 3: 

People management . This role pushes you to manage different types of teams and hold them together. Convincing the engineers to create a chatbot just the way I designed it requires strong people management skills where you can communicate your logic behind decisions without being assertive. 

Where do you work now? What problems do you solve? 

I work at Deloitte India

I solve issues faced by customers in different domains. For example, if it is Banking domain, I have to think about why users would come to a chatbot when they are banking, what will their intention be , what can the bot do to help them fast, how can I improve service through the bot. 

What skills are needed for your role? How did you acquire the skills?

Strong language skills, people management skills, problem solving mentality, pattern identification and leadership 

The main skills needed are language skills, psychology and mass communication skills. Having coding knowledge is good but not required. One needs to understand basics of communication and how software works.

What’s a typical day like? 

I participate in daily meetings to update ongoing activities, I design user journeys on flow chart makers, I write bot scripts, I think from user’s perspective and write full-fledged dialogues, I test designs, I clear doubts for developers and the testing team. 

What is it you love about this job? 

I get to see the end result. I can see how my design worked with the end users, I can learn new patterns, I can improve my work. This is the best part. 

How does your work benefit society? 

My work makes it easier for call center agents. The monotonous activities done by them are automated so that they can solve complex problems. I also help the users fix their issues, clarify doubts, finish tasks and simply keep them company if needed. 

Tell us an example of a specific memorable work you did that is very close to you! 

Most of the Bots that I design are close to me. Cannot think of anything specific. 

Your advice to students based on your experience? 

Be open to management activities. Participate in culturals and sports, focus on group activities. The people management and leadership skills that you acquire from school can help you travel a long way. 

Future Plans?

I live for the day. My future plan is to be happy and healthy. Anything that enables me to be that, I do.