Main Tech Enabler

We've seen some exciting possibilities

of how generative AI is being used to transform a variety of industries and

business tasks. As we're looking at them, hopefully you quickly realize that

these new possibilities seem to be far beyond the capabilities of a standard

GenAI chatbot that you're familiar with. So the key question for this course and

for this new technology is that how do we evolve this technology from a consumer

grade chatbot to a serious business asset, one that you can trust and manage as

integral part of your organization? The answer, as you might have realized, lies in one crucial aspect,

data, specifically, your data. To unlock the true potential of generative

AI to do things for your business, you must first connect it with

your business's unique data and let it learn from them to tailor its

abilities specifically for your needs. Let's see why this is so important by

revisiting the examples we discussed. In our first example,

you're using AI Copilot to organize and respond to emails on your behalf. However, you don't really want it to

write your emails like a generic chatbot, do you? No, you want the AI to craft your

messages in exactly the way you do. Now think about what the AI

needs to make it happen, to make your emails sound like you,

not ChatGPT. That's right. To achieve this, you need to train

the AI with your own past emails. You need to tell the AI all the tone,

context, priorities and nuances that are unique to

your communication style. Essentially, you need to find

a way to integrate a dataset of your previous communications

with the generative AI model. Only then can the AI learn from

your communication style and how you handle specific situations. Only after injecting a dose of your

communications into the model can it ensure that the emails it

generates feels authentically yours, saving you time while

maintaining your personal touch. In our second example, we talked

about how gen AI-based data analytics, which is one of the more promising

applications of this technology, to enable users to perform advanced

analytics using their own language as code, thereby mitigating the need for

all that programming. This could effectively make every

business professional a data scientist at the same time. Now think about what the AI needs,

though, to make it happen. This is a simple one. Well, to achieve this, you must be

able to efficiently import large and complex data sets into the model

before you can get started. Think about financial reports,

sales data, market trends, et cetera. You need to figure out a way

to effectively import or upload these huge variety of

complex datasets into the model, which is not an easy process, as raw data sometimes need to

be significantly processed and transformed first

before AI can understand their formats. We also talk about marketing, of using

GenAI to design marketing campaigns and simulating your target customers. What does the AI need to do

to acquire that capability? That's right, it needs your own unique

data on your past marketing campaigns and consumer engagements,

focus group results, et cetera, and the corresponding outcomes

of what worked for you and what didn't in these campaigns and

consumers. You need to inject these

data into the model first so it can learn from these interactions. Only by ingesting and analyzing these

historical data can the AI identify the key elements that contributed to your

successful campaigns and your failed ones. This learning is what enables the AI

to suggest optimized strategies for your future campaigns that are tailored

specifically to your target audience. We also looked at customer service. Now here you may ask, well,

isn't a chatbot designed for exactly that, to chat with your customers? Well, actually,

it takes quite a bit more than that. You don't want a bot to just

chit chat your customers, right? You want to provide direct, effective and personalized support that are specifically

tailored to the customer's current issues. So what does the AI need for

that to happen? Well, in this case, it needs access

to your history of existing customer interactions like chat logs,

order histories and support tickets. What do your customers

typically say in their chats? What are your agents' responses? Which responses got the best scores and

feedback from your customers? By giving it access to these unique data, you endow the AI with understanding

of your own customers issues, and therefore, create the best responses

that maximizes satisfaction. Only by arming your human

agents with these high quality, tailored responses generated by the AI

can they become more confident in focusing their time on handling

the more complex customer queries and therefore enhancing

the overall service efficiency. Moving even higher at a strategic level, GenAI needs even more data from your

business to assist in these functions. Take managing your supply chain risks, for

example, which is a hot topic nowadays. Sure, ChatGPT by yourself can

give you a beautiful report on the importance of managing

supply chain risks. That reads well. However, to unlock insights on how

to manage risks from your specific supply chains, again, you need to

inject your own risk data into it. For example, you need to supply first the

model with descriptions of risk sources that your business specifically faced,

any data on past mitigation efforts, and specific intelligence about

your suppliers, et cetera. Only then can you learn to predict

potential risk from these data and suggest the more accurate

mitigation measures for you. Hopefully, we can clearly now realize

one key thing about using GenAI for your business. It's not the AI

part that gives it value for your business, it is you. It is the power of your data that

gives GenAI its specific capabilities. The true power of generative AI for your business lies in its ability to

learn from your data from your business. Whether it's customizing communication,

performing advanced analytics, designing unique marketing campaigns,

enhancing customer service or managing risks, the key is to

provide the AI with relevant and specific knowledge from your organization. Only after you provide this

organization specific knowledge can this general purpose chatbot

be truly transformed into a specialized asset that delivers

value specific to your business. It is your data that gives AI the power. The technology merely give you a new and powerful tool to unlock value from them,

don't forget. Now let's deep dive into

an actual example of this. Let's see how a complex organization

with numerous stakeholders can successfully integrate generative AI

with their unique data and in doing so giving it additional

power to serve the organization.