AI in Action: Turning Theory into Reality. 26th February | 9:30 am
Jessica Chapplow
Cameron Goldie
Vinesh Chauhan
Time Section
9:00 am Guest Arrivals & Breakfast
9:30 am Welcome from Adobe
9:45 am AI: Turning Theory into reality – Jessica Chapplow
10:05 am Adobe presentation
10:35 am Break
11:00 am Panel Discussions
11:30 am Roundtable Discussions
12:00 pm Closing Remarks & Survey
12:05 pm Lunch & Networking
26th February 2025
How I ventured from AI theory into making Heartificial intelligence my reality Jan
Feb
Release of Stable LM 2
July
March
FTC announced new rules to protect people from fraudulent AI
August
OpenAI Sora Announcement April A
May
GPT-4o Omni released
June
Intel AI Chips Xeon 6 processor
Perplexity launches AI ad
In the form of a fake movie trailer during the NBA finals.
September
Claude 3.5 Sonnet
EU AI Act becomes binding law
OpenAI’s o1 model release
October A November
AI Pioneers win Nobel Peace Physics Prize
December
Amazon unveils Nova
AI models
DeepSeek-V3 launches
Source: PWC to the global economy by 2030 AI will generate 15.7trillion
Three waves of change are in focus Wave 1
Low-risk experimentation
Wave 2
Gradual adoption
Regulatory clarity
Wave 3
Lower barriers to entry
Widespread use cases
Proven reliability
Caution and uncertainty
Reshaped business and operating models
Meaningful role in decision making and strategic insights
The solution to the size and fit problem is knowing the personal preferences of your customers on each item-model-level.
UPS leverages ML to safeguard packages with DeliveryDefense. Enpal automated the generation of quotes for prospective solar panel customers. What’s happening with deepseek? Responsible AI Source: OECD
https://oecd.ai/en/ai-principles
Dimensions of responsible AI Dimension Example metric
Privacy & security
Fairness
Explainability
Robustness
Transparency
Governance
Is data used in accordance with privacy and legal considerations and protected from theft and exposure?
Are there harmful disparities in system performance across subpopulations?
Does the system offer a clear rationale for its decisions?
How hard is it to confuse or fool the system, for example, with “adversarial” examples?
Are users enabled to make informed choices about their use of the system?
How do you enforce and ensure these responsible AI practices are being carried out among all stakeholders?
Trade-offs in responsible AI 1. Dimensions of responsible AI can be at odds with primary objective to have best possible model performance
2. It is challenging to maximize all dimensions of responsible AI simultaneously
• Example: Making model results and components more transparent → potential security and privacy risk
• Example: Preserving privacy makes data more coarse → potentially degrades ability to explain model behavior
Why is responsible AI so complex? How do we make it a reality? Success is use case specific
Rapid changes in AI landscape (users, technologies, companies, societies; e.g. AI Act in EU)
Technical complexity (for example, root cause analysis of bias)
Simplify defining and measuring success; include bias metrics in evaluation
Improve continuously over the long term; adopt best practices now
Make responsible AI part of all DevOps/ML Lifecycle steps (next topic)
AI won’t outsmart humans, but those that know how to harness its full potential will. AI Use Cases in Personalised Journeys Dani Dennison
Deepen personalization beyond {First Name} Adobe’s AI capabilities will reshape every aspect of marketing; from planning, to creation, to execution
Adobe’s AI will be your teams’ co-pilot for transforming how content, campaigns, audiences, experiences, journeys and insights are generated at scale
Our approach to AI centers on scale, trust and enterprise readiness
Generative AI + Predictive AI: Transforming Customer Experience Management Enhanced Productivity
Insights-driven task automation
Intelligent Ideation
Ideation informed by past data
Optimized Journeys
Design for business outcomes
Strategic Innovation
Rapid data-driven experimentation
Hyper Personalization
Maximize on-brand engagement
Continuous Improvement Performance at scale
Create compelling, personalized, on-brand experiences AI-infused decisioning, experimentation, and optimization John, 26 Gold loyalty, frequent browsing, no recent purchase
End to end AI-powered marketer workflows End to end AI-powered marketer workflows What analysts are saying
“Adobehasbeen incrediblysuccessful thusfar.Itvery much paintedavisionforthe futurewithdeveloping acommerciallyviable serviceforgenerative AI…Ithinktheywill continuetosucceed”
“Adobeisinfusing generativeAIacrossits platformasit increasinglycatersto multiplepersonas rangingfromCFOsto CIOstoCMOsanddigital anddatachiefs.”
“Adobeisapproaching generativeAIina thoughtfulway, incorporatingbothits ownexperienceandthe collectivewisdomofthe broadercreative communitytodeliver somethingthat’s distinctiveandvaluable forallcustomer segments.”
“Asoftoday,Adobe Fireflyisthemost successful generativeAI productever launched.”
Source: Jay Pattissal, Forrester via The Straits Times:: ‘Make me an ad and find the customers’: Adobe rolls out AI tools for marketers March 2024
Source: Larry Dignan, Constellation Blog: GenAI Experience Cloud, customer Firefly models, marketing copilot with Microsoft, March 2024
Source: IDC via MarketWatch, March 2023
Source: Futurum Research, Mark Beccue
Research Note: Adobe Firefly: Blazing a Generative AI Application Trail October 2023
Solving content challenges lead to measurable returns Source: Adobe customer results via published case
Coffee Break Panel Discussions Jessica Chapplow
Tharishni Arumugam
Cameron Goldie
Vinesh Chauhan
Roundtable Discussions Please introduce yourself to the group.
How is your company leveraging/ planning to leverage AI to create personalised digital experience for your customers?
Can you provide an example of how AI has transformed a specific aspect of your company's digital experience? What were the key factors that contributed to this success?
What challenges have you faced from using AI in your customer experience strategy?