AI Mapping Canvas

AI is the new electricity.?

There was a time when electricity was a great new thing in the market, and it was a novelty to get the circuitry in the house and get the house all lighted up. With time various electricity related gadgets came up like fans, refrigerators, ovens, washing machines etc. It started to touch various aspects of our lives with various products and services. Now it’s all over the place in every walk of life. Just try imagining a world without electricity, everything will come to a standstill.?

In the evolving world of computing, AI is the new electricity. Everyday we are seeing it evolve and touch various aspects of all kinds of products and services. So, it is important that from time to time we map the various aspects of our products and services and discover and evaluate avenues of AI usage. This is important to keep the offerings?relevant and competitive in the marketplace. In the process of this mapping excercise, it also builds the knowledge, awareness and required AI skills of the teams working on these products and services. One of the ways to achieve this is to use “AI Mapping Canvas”.?

(ref: AI - Artificial Intelligence; HI - Human Intelligence)

Different Blocks of AI Mapping Canvas

Let us at look at some of the elements with example which will go in making of this canvas:

Key Features of Product & Services:?

For any given product or service, we start with listing various top level features. As we are listing them we also start to think from whatever we have known or heard about AI usage related to these features or what we think as the possibilities of AI usage. However it is important to understand what problems are addressed and the solutions offered. In this process, understand where all HI (Human Intelligence) is being used and can be replaced with AI (Artificial? Intelligence).? Eg, let us say an online store for selling books.? Couple of features could be like searching for books, previewing books etc. If we look at search features, it is filter based. Users will pick a category (say technology) and search for books of a given topic (say AI). However, given the power of conversational bots (eg. ChatGPT) we can have a more engaged and intelligent search offering. The bot can enable the search in a more conversational way. It can start with asking “what kind of AI books are you looking for?” and based on the answers it can suggest something or ask further questions or the user can prompt with further inputs to refine the search.? Imagine the old days, as we used to walk in a library or bookshop where the person there would help us locate the right books by having a conversation with us. We can bring the user closer to that experience using these bots. This is the first step towards discovery of such AI usage using the AI Mapping canvas.

Key Workflows and Usecases

The next step is to go a little deeper and start to list the key workflows and usecases. This involves going beyond the top level features. It involves going into various activities as clubbed together and using different features. When going through various activities it is very useful to think about the activities beyond just the digital offering and see from the point of view of real world physical scenarios where human interactions happen. Even in the digital offering to look out for where human intelligence is being used or could potentially be used. Eg. in online book store, one of the steps in the process of choosing a book is looking at the ratings and reviews. Now, when there are multiple reviews it could be quite time consuming to go over all the reviews. It would be great if it can summarize the gist of all the reviews. Also, at times we may be looking for specific information in the reviews (e.g. for an AI book, does it cover Generative AI, LLM, ChatGPT) and summarized in a nice way. Similarly ratings are generally covered in different categories. The customer may want to get a summary based on different categories and so on. These are interesting AI usecases which can come out of such an analysis of key workflows and functional usecases.

Key Data Sources & Analytics

There is a lot of data that gets generated as part of different features, workflows and usecases.? This step is all about understanding what all data is being generated, what are their sources? and how it is being used.? One of things to look out for is what part of this data is being processed in some automated manner to influence decisions. These decisions may be automated (handled programmatically) or assisting human intelligence in making some decisions. What more can be done to process the data to assist HI or augment automated decision making. Do we need to capture some new data which can further help us in this process? All this analysis is part of this step. Eg say, are there specific keywords for which the customers are summarizing the reviews for a given book. Can we prompt the customers with those keywords to provide feedback while editing and submitting the reviews.

Key Infra and Partners

For each of the above where is the data stored and how it is being handled from an infrastructure point of view. What data analysis tools are being used and why? What features do these tools provide and out of those what are we using? Do they provide AI tooling and are we leveraging them? Are there alternate tools available in the market? Can we use them or not? Reasons for not using them. What is the existing cost of the infrastructure and what can be optimized? What additional cost could be incurred on new tooling and revenue impact of the same. Eg. the SQL and No SQL databases that we may be using, do they come with some AI features? Can we leverage them? Is it part of our current payments (cost) system or we will incur additional cost. Will the additional cost help in driving additional revenue or better customer satisfaction? Eg. how will this work out, if we go for OpenAI ChatGPT APIs or some offerings from Hugging Face be a better option etc.

Customer Needs

The various features that have been built in the product and service, what are the customer needs that they are catering to? What are the different personas and their needs? What are different customer segments? What is the target group? Are there customer needs that we are not able to fulfill due to limitations of tech that we are using. Eg. in the case of an online book store there are a diverse set of customers. They may be looking for books related to tech, history, fiction, etc. Say, in case of tech, there could be times, the customer is not clear which book to go for. Like buying an AI book? But, for which level (beginner, advance)? Which domain (healthcare, security) etc. If we take along a person who is well versed with AI and go to a bookstore or take help of this person while searching online, then this person can give a lot of inputs in terms of choosing the right book. Can we bring in an intelligent conversational bot who understands the customer (given the various customer profile data available and data of similar customers) and assist the customer just like the human. That would be awesome. Also, if we can create some nice actual digital person which appears on screen and can converse just like a human on a video call (like a person standing next to us in a physical store).

Customer Options

While a given product and service may be able to cater to a lot of customer needs. However, there could be others which could also do the same and maybe more. What are the AI offerings they have? If we do not have them, then, that could drive away our customers. Also, what are the gaps that the competition has not yet filled and we can fill those. At times we may be able to create unique offerings which are hard to replicate or patentable and hence help us to establish our market position for a long period of time. With time the customer needs change. Eg. in an online book store slowly and gradually there is shift from physical books to eBooks. This opens up a whole new dimension. The kids may want their favorite story book read out aloud on an iPad in the voice and tone of their Mother. AI based solutions may be of help in these kind of solutions.

Customer Relationships

As things have moved online, so have the ways in which customer relationships are built and maintained. In the time when physical bookstores were the norm, the owner or staff at the shop were building and maintaining an in person one to one relationship with customers. However, with businesses moving online there is a lot of digital experience impacting the customer experience and relationship. We have more telephonic, email, chat etc. based communication that affects the relationship. E.g. in your favorite physical bookstore, majority of the time you will find your usual known person to be there to meet and greet you and assist you. Can this experience be generated with AI based digital personas? As soon as the landing page comes up, a digitally created human pops up and greets you and starts the conversation like “Hi, good to see you back? Did you like the AI book that you bought last time?”

Customer Acquisition?

The mode of customer acquisition has been changing with the changing time as the ways customers engage themselves and move around changes over time. There was a time when paper based newspapers were very popular and reaching customers through them was very powerful. Its marketshare got eaten by television and later television market share got eaten up by social media. With the coming of AI, the scope of personalization has taken a tremendous shift. In order to get better return on investment, it is important to evaluate AI usage in this direction. One of the other areas in which AI can be used is to analyze data and do predictive analysis of future trends and demands. This will help to get better prepared for the future.

Cost & Revenue

The impact of all the above on business in terms of financials is one of the most important aspects. To start with, we need to see what our existing costs and revenues are? Given the various AI usecases that we have come across through the above exercise we need to look into what investment will go in building them and what is the running cost of maintaining them. Next is to look into what kind of revenue impact this will have. A financial feasibility study needs to be done before proceeding further on it.

Based on the above , here is the summary of various different building blocks of AI Mapping Canvas:

• Key features of Product & Services
• Key Workflows and Usecases
• Key Data Sources & Analytics
• Key Infra and Partners
• Customer Needs
• Customer Options
• Customer Relationships
• Customer Acquisition?
• Cost & Revenue

As we do a walkthrough and dive into the above we come up with:

• AI usecases and their value proposition
• Additional cost and revenue impact and overall financial feasibility

In order to do all this you can use the AI Mapping Canvas which will look something like this :?

To work on this, you can either use an excel sheet for it or a white board or a Miro board.

It’s best to do it as a whole team with the help of a workshop coach. Coach helps to moderate the brainstorming session and maintains the rhythm and structure of the workshop. Also, the coach will bring in the discussion the various elements of AI usage and technologies available which can be relevant to the discovery process.

If you are interested further in knowing about this workshop, please message me on LinkedIn or leave a comment in the section below and will reach out to you.

It's time to get AI ready! Go for it!