Ce diaporama a bien été signalé.
Le téléchargement de votre SlideShare est en cours. ×

A brief history of artificial intelligence for business

Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité
Publicité

Consultez-les par la suite

1 sur 29 Publicité

A brief history of artificial intelligence for business

Télécharger pour lire hors ligne

Since the 1960s, Artificial Intelligence has promised us benefits in business and in our personal lives. This presentation takes us from the early days up to machine learning and applications for enterprise businesses that are delivering personalized experiences to customers ... to a "segment of one."

Since the 1960s, Artificial Intelligence has promised us benefits in business and in our personal lives. This presentation takes us from the early days up to machine learning and applications for enterprise businesses that are delivering personalized experiences to customers ... to a "segment of one."

Publicité
Publicité

Plus De Contenu Connexe

Diaporamas pour vous (20)

Les utilisateurs ont également aimé (20)

Publicité

Similaire à A brief history of artificial intelligence for business (20)

Publicité

Plus récents (20)

A brief history of artificial intelligence for business

  1. 1. Copyright © 2017, Datalog, Inc., all rights reserved. All trademarks property of their respective holders ® A brief history of artificial intelligence for business Jack Crawford, Founder and CEO
  2. 2. Arthur L. Samuel, 1959 2
  3. 3. Back then, AI was a forest of trees citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.368.2254&rep=rep1&type=pdf
  4. 4. In the 1970s interest in AI renewed Artificial intelligence attracted money like bears to honey 4Image: zooportraits.com VC Bear
  5. 5. and the money dried up Then, the bears got burned 5Image: The Huffington Post
  6. 6. After a long AI Winter, the winds of Δ ignited interest again 6Image: The BBC, “AI: 15 key moments in the story of artificial intelligence” The first money saving business AI was an “expert system” built at DEC in the 1980s
  7. 7. In the 1990s, other businesses began to to employ “expert systems” 7Illustration: learnlearn.co.uk EXPERT SYSTEM
  8. 8. It was the age of … 8 Rules
  9. 9. Rules to put people (us) into buckets 9 Market Segmentation
  10. 10. Then Jeff Berry had a bright idea 10 “Each customer should be seen as a SEGMENT OF ONE Image: LoyaltyOne
  11. 11. 11Image: Getty Images, Cristian Baitg
  12. 12. Which ushered in the age of the … 12 Individual
  13. 13. The problem was that at that time … 13 The only way to predict individual behavior and act on it in a timely manner involved using rules Image: Pegasystems, “Next Best Action,” youtu.be/HeL-Y1kSoDg
  14. 14. RULES 14
  15. 15. Maybe, just maybe … 15 We don’t need rules
  16. 16. Two years ago at the 29th AAAI 16 aaai.org/ocs/index.php/AAAI/AAAI15/paper/viewFile/9444/9488
  17. 17. Here’s a true story The problem 1.Our client had 49% market share YOY 2.It dropped to 46% in the past year 3.Customer switch was the culprit The goal Find out which customers were likely to switch That is, identify the segment of 1 17
  18. 18. Steps that we took 1. Ask the CMO if we could try 2. Listen to her laugh and say “Sure. Good luck with that.” 3. Call a meeting with the data guardians, citing approval from their chief marketing officer 4. Wait to receive the data we requested (1 month) 5. Load and clean up the data (1 hour) 6. Run the model on 70% of the data (1 day) 7. Verify prediction with the remaining data (1 day) 8. Spend time improving results so they don’t think it was easy (1 month) 9. Give them the test results (92% accuracy over the past 3 years) 18
  19. 19. So what did we hope they would do with this miraculous information? 1. Create strategies to identify interventions for customers with high propensity to switch to their competitor 2. Establish “A/B” testing (or other measures) to validate the effectiveness of these strategies in practice 3. Put these interventions into the “field” 4. Realize the benefits through recovery of their market share 19
  20. 20. This particular story had a surprising twist, and led to our AI startup They ”shelved” the idea Why? Other market factors might have caused the switch 1. Slow pace of product improvement 2. Reduced advertising budgets 3. Sales force effectiveness gaps 20
  21. 21. And that “made sense” in 2015? 21 At the time, many businesses couldn’t see the power of machine learning for predicting the buying behavior of an individual consumer
  22. 22. It seems that many innovations take forever to be adopted by business – 2009 22machinelearning.org/archive/icml2009/papers/218.pdf NVIDIA GTX 280
  23. 23. You can make AI valuable for business – opportunities Sales and marketing Next best action Salesrep coaching Customer switch Supply chain Supplier commitment prediction Conversational bots Customer care Propensity to call Conversational AI Personalization Research and development Knowledge retention Success factor identification 23
  24. 24. Beyond predictive AI, much more opportunity lies ahead 24
  25. 25. The next leap in AI is 25 Conversation
  26. 26. MyPolly.ai ™ Join our conversational AI beta at:
  27. 27. 27 postscript: Let’s remember the pioneers who began the journey that brought us here today
  28. 28. The reunion of some early AI researchers 28Image: The New York Times, December 7, 2009
  29. 29. Copyright © 2017, Datalog, Inc., all rights reserved. All trademarks property of their respective holders ® Thank you 29 949 • 374 • 1120 Jack@datalog.ai ™

×