Day 152 MIT Sloan Fellows Class 2023, AI for business 4 "Framework for AI business"
AI Framework
New AI business idea with framework
Presentation design is pretty personal and cultural. It depends on your preference about design, complexity of messages and organizational culture.
Even with completely same message, the best slide design to deliver is different.
Problem Statement
If from people's feedback about presentation designs, we could know the best presentation format for specific messages, audience, and context, then many young presentors could derive a benefit by reducing time to modify presentation design and focusing on messaging/contents.
So many business person recreate the similar deck with similar message from scratch. The organizations such as my company, an agency, should store the pitch data in the standardized way and then we can analyze and optimize presentation design in the data-driven way.
Inputs(x)
- Audience attributes, recency, contexts/situation, messages and design of presentation(Image data)
- Recorded video of presentation might be future training data.
- Contexts and messages are difficult to be standardized.
- This would cause some misinterpretation of contexts.
Outputs(Y)
- Pitch feedback and people's feedback (score or Y/N) / budget from pitch
- The past result would affect the future result. (The same message with same design presentation would give a negative image for clients even if it recorded positive score before)
- Scarce patterns of presentations would be a noise for majority of cases.
- Major concern is too many patterns of designs and contexts.
Training data
- Actual recording video from presentations
- Pitch results and presentations
- We don't have enough data so far. Data storing project need to be initiated.
- The presentation data is basically owned by agencies, but it may contain some confidential data for clients.
Errors
- It may cause wrong suggestions to future presentation.
- Also, it would cause the bias in the future training data.
