0% found this document useful (0 votes)
73 views10 pages

Artificial Intelligence

The document provides templates for building an AI product blueprint across 7 layers: 1) defining the problem, 2) monetization approach, 3) key questions, 4) user interface, 5) machine learning architecture, 6) important features, and 7) available data sources. The overall goal is to connect business outcomes to technical outputs to solve customer problems.

Uploaded by

Muni Yugandhar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
73 views10 pages

Artificial Intelligence

The document provides templates for building an AI product blueprint across 7 layers: 1) defining the problem, 2) monetization approach, 3) key questions, 4) user interface, 5) machine learning architecture, 6) important features, and 7) available data sources. The overall goal is to connect business outcomes to technical outputs to solve customer problems.

Uploaded by

Muni Yugandhar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

How to cross this Chasm ?

An actionable playbook to connect 3 dimensions

OUTPUTS OUTCOMES
( “Math” ) ( “Business” )
7 Layers to Peel for AI Product Blueprint
1 Problem worth solving
2 Monetization 1 liner
3 PUQ
4 Screen Interaction experience
5 ML Architecture
6
Feature Vectors
7
Data Pools
AI Product Template-1 :
“KAAN” Problem Crystallisation Worksheet
What's the one problem we are solving ?
• “My unsolved problem is __________________________________________________.

• This is keeping me awake at night and it is important to solve this problem because
___________________________________________________.

• If I don't solve this problem I am going to have a __________________________________________________.

• We need to improve L2B from __________________________________________________.


• to ___________________________
• In order to improve L2B we need to do following
• ___________________________
• ___________________________
• Current issues are
• ___________________________
• If I solve this problem the BEST CASE $ impact is ____USD ___________________________
• WORST CASE $ impact is ____USD ___________________________

• This is among the top _____ problems I need to solve this year”

cc: VinothChandar - https://www.flickr.com/photos/44345361@N06


AI Product Template-2 :
AI monetization 1 Line Conversation starter
How do you think analytics can unlock value ?

Please write down your high level estimated model …

It need not be perfect … Intent is to spark a conversation


AI Product Template-3 :

PUQ
What are your top 10 “Powerful Unanswered Questions” ? Core blind spots

4 Pulse blindspots 3 Forensic blindspots 3 Prediction blindspots


Situational Awareness of “Vitals” Insights Prediction
“What's happening here ?” “Why ? Why ? Why is it happening?” “What's likely to happen here ?”
1 1. 1.

2 2. 2.

3 3. 3

4
AI Product Template-4 :
PROTOTYPE
Can you sketch the landing page of the AI product the way you envisioned it ?

Your company logo here


1. Logo 2. <AI Product Name >
: :
Your company AI Product name here

Input What are the inputs to AI product ?


1. Customer id : - Customer id ?
- Product id ?
- Region ?
- Asset id ?
- Process id ?

Output
Interaction experience
1. Top 3 products to recommend
What are the Outputs from AI product ?
- Scores
- Next Best Action
- Propensities
- Graphs ( Line, Bar, Pie, Heatmaps)

6
AI Product Template-5 :
MACHINE LEARNING CONSTRUCT
Which ML/DL technique can solve your business problem ?

1. Collaborative Filtering ?
2. Text Mining ?
3. Graph Analysis ? Queen bee, Community detectors
4. Scoring Models ?
5. Sequence analysis ? Markov chains
6. Optimisation Models ?
7. Ensemble Models ?
8. LSTM Models ? 4 flavors
9. CNN Models ?
10. Bot Assistants using NLTK
11. Clustering
AI Product Template-6 :
FEATURE VECTORS
Which feature vectors must we curate ?

Recency vectors Frequency Intra event Dispersion Velocity vectors Volatility Normalized
vectors vectors vectors vectors vectors

 Zero results returned  Mobile visit  Intra login interval  Payment failure  Premium purchase  Relative spend index
frequency dispersions velocity
 Late night logins  Price insensitive  Intra shopping  Search dispersions
searches interval
 Location dispersion
AI Product Template-7 :
MAP INFORMATION LANDSCAPE
Where does the data live ? Data pools ?

1. 2. 3. 4.
Put data pool name here Put data pool name here Put data pool name here Put data pool name here

9
Putting it all together - Blueprint
1. Problem worth solving Reduce Churn by 3 % in 12 months

Best Case = $2,000,000


2. Monetization 1 liner Realistic Case = 1,800,000
Worst case = 1,500,000

3. PUQ

4. Screen

5. ML Architecture

1. Visit frequency
6. Feature Vectors 2.
3.
Intra visit interval
Average spend dispersion

7. Data Pools

You might also like