270 likes | 278 Views
WEBINAR The Promise And Potential Peril Of AI. Martha Bennett, Principal Analyst Matthew Guarini, Vice President, Research Director. March 15, 2017. Call in at 10:55 a.m. Eastern Time. Agenda. Introduction: Why is AI such a hot topic today? How does Forrester define “AI”?
E N D
WEBINARThe Promise And Potential Peril Of AI Martha Bennett, Principal Analyst Matthew Guarini, Vice President, Research Director March 15, 2017. Call in at 10:55 a.m. Eastern Time
Agenda • Introduction: Why is AI such a hot topic today? • How does Forrester define “AI”? • What are some of the key application areas for AI? • Which are the main pitfalls you should avoid? • How should you drive your organization’s AI strategy?
Artificial intelligence (AI) isn’t one single technology • Like “cloud” and “big data,” it means different things to different people and consists of a wide variety of different technologies. • The technologies under the AI umbrella are at different stages of maturity. • Source: TechRadar™: Artificial Intelligence Technologies And Solutions, Q1 2017 Forrester report
Examples of pragmatic AI include: • Speech recognition and natural language processing and generation • Alexa, Siri, Google Assistant, and Cortana • Chatbots on websites and in mobile apps • Automated translation • Machine learning and knowledge engineering • Lots and lots of application areas, ranging from reducing churn, predicting propensity to buy, personalization on the fly, recommendations, fraud detection, predictive maintenance, etc. • Image recognition combined with machine learning and deep learning • Detecting microscopic flaws and differences in manufactured goods • Prioritizing brain scans for examination by a clinician
Examples of pragmatic AI, continued: • Advanced discovery techniques: • Machines are far superior to humans when it comes to data crunching — try reading millions of pages of specialist literature in seconds . . . • Flagship example: IBM Watson for Oncology • Robotics and self-driving cars • Self-driving cars aren’t about to take over, even though they’ve already proven to have lower accident rates than humans. • Robotics isn’t just about machines taking over jobs — a key application area is supporting human workers.
True “cognitive” systems are as far away as ever, but companies can leverage AI technologies today. Source: Artificial Intelligence: What's Possible For Enterprises In 2017 Forrester report
Interest in AI is high, adoption is in the early stages Source: The Promise And Potential Peril Of AI Forrester report
Q: “How much has your team done to date with AI?” Poll Actively using Implementing Piloting Studying Nothing
Q: “What areas of your organization are leading or evaluating the investment and adoption in AI systems?” Current AI investment is customer-focused. Base: 418 business and tech professionals; “don't know” answers excluded from analysis; Source: Forrester’s Q2 2016 Global State Of Artificial Intelligence Online Survey
Q: “In areas where your organization has piloted, tested a POC, or deployed AI, what results have you achieved?” Early adopters are seeing results. Base: 394 business and tech professionals; “don't know” answers excluded from analysis; Source: Forrester’s Q2 2016 Global State Of Artificial Intelligence Online Survey
Many AI scenarios revolve around augmenting human expertise, not replacing it.
There are times when machines perform better than humans . . .
. . . but there are also many scenarios where humans and robots enhance each other.
Robots can help address labor shortages. Head chef: robot Andrew flips a pancake in the Henn-na restaurant at the Huis Ten Bosch amusement park in Sasebo, Nagasaki — The Asahi Shimbun/Getty Images Source: Financial Times, February 1, 2017
But it’s not all brightness and sunshine Key pitfalls include: • Training bias and overfitting. • Use of inappropriate techniques. • Underestimating complexity and skills requirements. • Not considering how a system could be subverted. • Compliance/ethics violations.
But it’s not all brightness and sunshine Key pitfalls include: • Training bias and overfitting. • Use of inappropriate techniques. • Underestimating complexity and skills requirements. • Not considering how a system could be subverted. • Compliance/ethics violations. AI isn’t magic — it’s a lot of hard work!
Recommendations • Drive your organization’s AI strategy — how can it benefit? • If you don’t, your competitors probably are. • Develop a road map for AI — start small, but think big. • Incremental benefits or business transformation? • Build capability in your team to be the go-to-resource of AI. • Otherwise, business decision makers will go elsewhere. • Make sure you take an end-to-end process view, and also look across processes. • Beware of unintended consequences. • Start the conversation about legal and ethical issues at board level. • Your organizations want to be in the headlines for the right reasons!
Further reading • The Promise And Potential Peril Of AI Forrester report • TechRadar™: Artificial Intelligence Technologies, Q1 2017 Forrester report • Artificial Intelligence: What’s Possible For Enterprises In 2017 Forrester report • Artificial Intelligence Revitalizes BPM Forrester report
Forrester insights for iPhone and iPad KEY RESEARCH AND DATA POINTS WHEN AND WHERE YOU NEED THEM • Access playbooks, reports, key takeaways, and data points to accelerate your projects and support your decision making. • Save reports and graphics to read online or offline, on the device of your choice. • Receive notifications to stay abreast of the latest trends and insights relevant to your initiatives. forrester.com/app
Martha Bennett +44 7768 896 540 mbennett@forrester.com Twitter: @martha_bennett Matthew Guarini +1 630-862-4940 mguarini@forrester.com Twitter: @guarini12