Snowforce how to drive rapid business value with ai
- 1. How to Drive Rapid Business Value with Artificial Intelligence Philip Cooper VP, Product Go-to-market Salesforce Einstein
- 2. AI Assists Us With the Little Things Before Work On the Go At Home Google, play my morning playlist. Playing “Morning Pump Up” playlist..
- 3. The Same is True for the Enterprise AI has immense potential to assist us at work Analytics “Do any of my accounts have whitespace?” “How do I make my service agents more productive?” ”How do I improve the yield on my campaign spend?” AI-Powered Apps “Which leads should I be focusing on?” “Which products should I recommend to my customers?” “How do I prioritize and route my cases?” Natural Language Processing “How do I leverage my unstructured data?” “How do I predict customer sentiment?” “Can I learn anything about my customers from text and images?” Conversational UI “How do I automate routine tasks for my employees?” “How can I deflect straightforward customer requests?” “How do I increase revenue from my website?”
- 4. of enterprise execs believe AI will help them create a competitive advantage of the same companies have an AI strategy 84% 38% Source: “Reshaping Business with Artificial Intelligence” MIT Sloan Management Review & Boston Consulting Group Executive Imperative
- 5. Economic Opportunity 70% of companies will adopt at least one AI technology by 2030 delivering: Global GDP boost of 1.2% per year $13 trillion of additional economic output Source: McKinsey Global Institute
- 6. Hurdles in Implementing & Adopting AI Remain 47% AI embedded in at least 1 business unit/function Up from 20% in 2017 21% AI embedded in >1 business units/functions Only 1-in-5 are “Pioneers” These companies are investing in Revenue- generating applications of AI and 9 out of 10 have a strategy “Paradox of AI” AI pilots are easy; Success at scale is much harder
- 7. Adoption Varies Widely by Industry & Function Source: McKinsey Global Institute
- 9. AI Artificial intelligence: making machines exhibit humanlike intelligence Smart robotics, autonomous vehicles, virtual agents, machine learning Machine Learning Models that learn from historical data to predict outcomes Spotify, Netflix, Amazon Deep Learning Complex algorithms built to mimic the way the human brain learns Self-driving cars, computer vision, audio recognition, machine translation Natural Language Processing Training computers to understand, interpret and manipulate human language Siri, text analytics, speech recognition, natural-language understanding, natural-language generation In the Absence of a Common Definition…..
- 10. AI-Powered Solutions are Special
- 11. Scale What Makes AI Different? Complexity Opacity Bias Change
- 12. Maximizing the Adoption & Impact of AI Engage the Business Jointly identify use cases Collaborate via a common language Educate your end-users Set Achievable Goals Start simple Ensure measurable performance Approach Deliberately Start small Experiment & learn Follow an agile process Measure Impact Create baseline analytics Track model accuracy Measure lift / gains Eric Chaniot Chief Digital Officer Michelin Invest in UI/UX Consider explainability Deploy to existing workflows Put results into business contexts Bake in Ethics Foster an ethical culture Build transparent models Scrutinize training data
- 13. What are the Key Enablers of Success? DATA CHANGE TRUST
- 14. Where are my data stored? Key Considerations for Successful Enterprise AI: Data Do I have equitable training data? Do I have the right data for my use case?
- 15. Are the right internal stakeholders involved? Key Considerations for Successful Enterprise AI: Change Am I drastically changing an existing workflow? Can I prove ROI for my end users?
- 16. Am I embedding AI in existing systems & workflows? Key Considerations for Successful Enterprise AI: Trust Am I baking user feedback into my models? Am I providing model transparency to my users?
- 17. 3 Ways to Build Ethics Into AI
- 18. Key Enablers of Success DATA CHANGE TRUST
- 19. Get Started! 1. Select your use cases 4. Deploy & integrate into Business 2. Evaluate Data Readiness 5. Measure Impact 3. Prep Data and Train Models
- 20. Audience Q&A
Editor's Notes
- #2: Title: How to Drive Rapid Business Value with Artificial Intelligence Abstract: Artificial Intelligence is the hottest and yet most daunting set of technologies facing enterprises today. There is immense opportunity to utilize AI to affect business outcomes, but how do you get started quickly and effectively? In this session, Phil Cooper, VP, Product Go-to-Market for Salesforce Einstein, will discuss key considerations in getting started on your journey with AI, delivering successfully and driving mind share and adoption along the way.
- #3: AI already assists us in the little things that make our lives smoother every single day. From Google Home helping us get our day started in the mornings with our favorite playlist To Waze helping us avoid traffic on our way to work To Starbucks recommending which seasonal coffee drinks I should order based on our past orders. But these are all the ways that AI assists us with the little things outside of work. When we step inside our offices we don’t have the same level of AI assistance. What if you could bring the power of an AI assistant into your work life and change the way you work every day?
- #5: - Fall 2017 - 84% of enterprise execs know that AI will give them a competitive advantage and only 38% of the same companies have an AI strategy
- #6: September 2018 – modeling impact of AI https://www.mckinsey.com/~/media/McKinsey/Featured%20Insights/Artificial%20Intelligence/Notes%20from%20the%20frontier%20Modeling%20the%20impact%20of%20AI%20on%20the%20world%20economy/MGI-Notes-from-the-AI-frontier-Modeling-the-impact-of-AI-on-the-world-economy-September-2018.ashx
- #7: Progress…BUT McKinsey Survey Nov 2018 https://www.mckinsey.com/featured-insights/artificial-intelligence/ai-adoption-advances-but-foundational-barriers-remain Capabilities most commonly deployed: robotic process automation computer vision machine learning are Industries: telcom, high-tech, Fins, PS Functions: Service, Product, Marketing & Sales BCG / MIT Sloan 3,076 companies surveyed 1/5 are in the Pioneer category (Pioneers, Investigators, Experimenters, Passives) Companies are moving from productivity gain applications to revenue-generating applications Only 38% of CXOs see AI reducing employment China is moving fast and companies are determined: business models are being adapted to take advantage Early investing are doubling down (not pulling back and reaping the benefits)
- #10: …..not including robotics, autonomous vehicles etc.
- #12: Change is a constant Separate solutions interact and become entangled
- #15: where are the data? are they accessible? Are data aligned with my use cases? Is there signal in my data? Are there enough data? Are the data available at runtime for scoring? Is there actionability in my data? are my data equitable? do I need to worry about data quality?
- #16: Are the right internal stakeholders involved? Are LOB executives educated on AI? Am I embedding AI in existing systems & workflows? Does my business process or my user workflow need to change? Can I prove ROI for my sponsor?
- #17: Am I providing model and scoring transparency to my users? Am I proving ROI / impact for my end user? Am I taking user feedback into my models?