Skip to Content

Transformation Trends: Highlights in Artificial Intelligence From Dreamforce ’19

parker harris and marc benioff announce new Einstein features onstage at dreamforce ’19

Artificial intelligence’s potential to help us work smarter is immeasurable. Still in its infancy, we’re collectively learning how to use it, regulate it, and build it responsibly.

Artificial intelligence (AI) is changing the way we interact with the world. Questions about AI and its implications were heard all over Dreamforce ’19:

  • What is AI?” 
  • “Should I be worried about AI?”
  • “AI isn’t applicable to me … or is it?”
  • “I want AI to help me with my job. How do I get it?”

As Co-Founder and CEO of pymetrics Dr. Frida Polli joked, “AI is like teenage sex: everyone says they’re doing it, but no one really knows what it is.” In multiple keynotes spanning the event’s four days, AI experts shared their insights and advice to address some of the confusion. Here’s what we gleaned:

Marc Benioff einstein speaker
Salesforce Chairman and Co-CEO Marc Benioff shows the audience an Einstein smart speaker during Dreamforce’s opening keynote.

Take action with your voice

“The future of AI is conversational,” said Salesforce Research’s Caiming Xiong before showcasing a fully-autonomous AI customer service bot receiving and processing a pizza order. Behind him, a live logic decision-tree graphic showed the audience in real-time how it worked. With voice technology business use on the rise, we’re seeing how companies interact with customers transform. According to the Enterprise Technology Trends report, while only 36% of enterprise IT organizations use customer-facing voice technologies today, another 50% plan to do so within the next two years. Here’s why: by the end of 2019, it’s expected that smart speakers’ global installed base will top 200 million. Even broader, the market for voice assistants powering smart speakers is on track to reach 8 billion by 2023.

The B2C world has jumped on this trend thanks to voice skills platforms like Amazon’s Alexa and Google Assistant. These two large players have made it easy for both experienced developers and people without coding backgrounds to explore voice’s potential via skill kits, tutorials, training courses, and more. For example, with a simple vocal command you can now use Alexa to send or request money through PayPal, prep for a night out by buying movie tickets via Fandango, or stay in and cue up a movie on Netflix. Imagine leaving home on a chilly morning to have your car’s engine already warmed up, door unlocked, interior temperature at a comfy level, with your commute playlist cued-up. Car manufacturers like Ford, Hyundai, and Toyota have made this scenario a reality, allowing drivers to interact with their cars from the comfort of their living rooms, hands-free.

That’s not to say that B2B companies haven’t kept up; businesses are increasing voice technology usage to interact with customers and improve operations. For example, Vistaprint offers up small business marketing advice, included in the relay of your daily news feed. Companies like WeWork and Conde Nast have created voice-enabled conference rooms where employees can manage meetings and control conferencing systems. Recently, Marriott rolled out Alexa for Hospitality in select hotels, making hotel services (e.g., housekeeping, concierge, front desk, and back-office systems) accessible through voice interaction. Last year, we introduced Einstein Voice to help sales teams verbally log updates in Salesforce through our mobile app. At this year’s Dreamforce, we expanded on that with Einstein Voice Skills to enable every company to build custom, voice-powered apps tailored to any role or industry.

As natural language processing technology advances, we can expect this paradigm shift towards voice as a standard service and operating tool. This shift means improving the business capabilities of voice assistants, chatbots, sentiment analysis, and machine translation. Since voice is the most natural way for customers to communicate, we’ll see increased adoption of the technology to improve customer experiences.

AI implications
(L-R) Kathy Baxter, Dr. Vivienne Ming, and Kay Firth-Butterfield discuss the ethical implications of AI.

AI ethics is a mindset, not a checklist

According to Socos Labs Co-Founder and Executive Chair Dr. Vivienne Ming, ethics is about making hard decisions when your interests and society’s needs diverge. Ensuring ethical AI starts with asking the right questions like, “What needs to be solved” and “Is this something people want or need?” Ming advises that you may not necessarily like the answers to these questions. “[Ask yourself,] am I making these decisions with my own self-interests in mind or others’ interests?” she says. For example, deciding to build an algorithm fast, or slowing down the process to ensure the end-product doesn’t harm people.

Ming shares these tips because the AI community is now largely self-regulating. With the world in various states of AI governance, it’s up to the companies and individuals developing and creating AI products to think through ethical concerns to build responsible solutions.

And these concerns from data scientists and the general public aren’t unfounded. Tech industry leaders like Microsoft’s Bill Gates and IBM’s Ginny Rometty agree that the boundless potential of AI could affect global economies and societal systems — for better or worse. For instance, on a large scale, AI technologies are already being used in military operations, global financial institutions, and healthcare systems. Closer to home, there are apprehensions about the effects of AI including automation, unemployment, and the cause of foundational inequalities.

The danger of inadvertently creating inequalities was top of mind for AI panelists at Dreamforce. Creating and perpetuating bias in AI is a serious concern for the AI community because when used carelessly, AI magnifies the human-created stereotypes, biases, and false information that already exists — which, in turn, influences decision-making. We’ve seen it in multiple examples from job screenings to mistranslations on social media to racial profiling in the criminal justice system.

To combat bias and prioritize ethical AI, the World Economic Forum’s Head of AI and Machine Learning Kay Firth-Butterfield, stressed putting ethical considerations into practice throughout the entire lifecycle of product development, from conception, to design, development, and sales. She also advises that the business world needs to carefully consider where and when to implement AI so that it helps us overcome our limitations. “We really need to think about if we need AI everywhere, and making sure that AI is lifting us up as humans, rather than oppressing us,” Firth-Butterfield said.

(Interested in learning more about combating bias in AI? Read up on how the Salesforce Research team is building AI tools the right way, and learn more on Trailhead)

How AI will affect our workplace in the future

(L-R) Edward Felsenthal, Lili Gangas, and Dr. Frida Polli talk about how AI will affect our workplace in the future.

Make reskilling a priority

In 2016, the World Economic Forum forecasted that nearly 35% of the skills required for jobs will change across industries by 2020. They also noted that at least one in four workers in OECD countries already report a mismatch with regards to the skills demanded by their current job. Although many companies want to help workers reskill and prepare for the massive transformation that AI bring, they’re finding it too difficult or too expensive to focus on. Yet, the conversation at Dreamforce emphasized that it would be a costly move to ignore this problem as AI’s growth outpaces the number of skilled workers available.

As Kapor Center Chief Technology Community Officer Lili Gangas shared, “This is going to be a $15 trillion industry, but when there is nobody who looks like me [a woman of color], my mom, or relatives at the table, we’re really missing a huge chunk of the potential talent.” To create the workplace of the future, leaders must prioritize reskilling (perhaps with a dedicated corporate role) and tap into new, more diverse talent pools.

We’ve seen a correlation between companies committing to diversity and having more success. Mckinsey finds companies that are in the top quartile for both racial and gender diversity are more likely to have financial returns above their respective national industry medians, 35% and 15%, respectively. Accordingly, a diverse workforce benefits from having multiple perspectives to address and minimize AI bias and create ethically sound products. Dr. Frieda Polli compares AI to toddlers learning from grownups and mentions that most of the people working in AI-related fields are young men in their 20s. “Imagine if all the world’s toddlers were being raised by these men,” she quips.

But how can companies take steps to reskill their employees and diversify workers? Polli suggests looking at current hiring systems, which is based on resumes, doesn’t scale, often overlooks soft skills and human potential, and bypasses underrepresented groups. Businesses can take a process-based approach to solve this, including the practice of blinding resumes (removing non-job-related information) and looking beyond traditional university programs to other kinds of workforce training programs.

Companies can also prioritize reskilling by:

  • Fostering a continuous culture of learning. Helping people to update their skills in the fast-moving AI-age will be an ongoing process. Innovative companies look beyond the traditional educational platforms to retrain and reskill existing workforces.
  • Democratizing AI. Companies are already experimenting with new user interfaces including voice-technologies and robotics. Training programs should ensure that users know how to interact with AI tools and systems.
  • Enabling data literacy. As data fuels AI systems, ensuring and promoting a data literacy culture is essential.
  • Making software easier to use. Software that is customer-centric, broadens the participation of people who can use it. With more configurable, drag-and-drop-type interfaces, more people can contribute to development.

An optimistic future

Artificial intelligence’s potential to help us work smarter is immeasurable. As this technology is still in its infancy, we’re collectively still learning how to use it, regulate it, and build it responsibly. While many people still fear and worry about the repercussions of new technology, our experts at Dreamforce all share a similar mindset: optimism.

They, like us, are excited about the future and to see how we’ll collectively tackle the challenges and create opportunities in our changing societies. Gangas shared, “I see middle and high school girls who are creating algorithms for autonomous vehicles — for fun! So, I’m really optimistic about the future that we have. We want to make sure we’re creating equitable pathways for them to join us.”

Want to learn more about AI at Dreamforce? Read our blog post, “Why We’re Bringing Voice to CRM” or check out the handy video playlist on Salesforce LIVE. Alternatively, you can jump right in and get started with our AI Trailmix on Trailhead, our fun, free online learning platform.

 

Kim Mercado Associate Manager, Editorial Content

Kim Mercado was a writer for the Salesforce blog. She's written many of your favorite blog posts here, as well as multiple e-books including the Content Marketing Institute award-winner, We Are All Trailblazers. When she isn't writing, she's probably binge-watching tv shows, learning Japanese, and looking at videos of cute animals.

More by Kim

Get the latest articles in your inbox.