salesforce research

Top Generative AI Statistics for 2024

10 min read

Key Takeaways

Generative AI has unleashed exciting new opportunities for workers across IT, marketing, sales, and service. Salesforce’s generative AI research, highlighted in this article, reveals how these roles are using the technology, why they’re not, and much more to help businesses prepare for the future.

Editor’s Note: This story was updated in September 2023 to keep it fresh.

Generative artificial intelligence is on nearly everyone’s mind – and Salesforce has the stats to prove it.

In addition to Salesforce executives giving a primer on generative AI and sharing predictions about its effect on the future of sales , small businesses , and CRM , the company has spent months researching trends and worker sentiment about the new technology.

Salesforce’s Generative AI Snapshot Research Series , an ongoing study of over 4,000 full-time workers across industries, reveals the power of and perspectives on generative AI across IT, marketing, sales, and service – and an additional study among over 4,000 general population adults across the U.S., UK, Australia, and India.

For a comprehensive look at industry and generative AI research, visit the Salesforce Stat Library . For industry-specific insights on generative AI, click the relevant section below:

General population data reveals new insights

Salesforce’s most recent survey on generative AI use among the general population within the U.S., UK, Australia and India found the public is split between users and non-users. Within each country, the online populations surveyed reported the below usage (note: cultural bias may impact results):

73% of the Indian population surveyed is using generative AI.
49% of the Australian population surveyed is using generative AI.
45% of the US population surveyed is using generative AI.
29% of the UK population surveyed is using generative AI.

Notably, the research uncovered AI users are “super-users” – meaning the majority of them use the technology frequently and believe they are well on their way to mastering it.

65% of generative AI users are Millennials or Gen Z, and 72% are employed.
Nearly 6 in 10 users believe they are on their way to mastering the technology.
70% of Gen Z report using the technology and 52% of them trust the technology to help them make informed decisions.
52% say they use generative AI more now than when they first started.

These early adopters are also looking to expand generative AI use cases from fun and games to workplace and enterprise applications:

75% of generative AI users are looking to automate tasks at work and use generative AI for work communications.
38% of generative AI users use generative AI for fun/messing around, and 34% for learning about topics that interest them.

Not everyone is sold on AI just yet – and it tends to fall along generational lines. While AI users are young and engaged, non-AI users are older, less familiar with the technology, and more skeptical of its benefits:

68% of non-users are Gen X or Baby Boomers.
88% are unclear about how generative AI will impact their life.
40% say they aren’t familiar enough with the technology and 32% say the technology is not useful for them.

More cautious than early adopters, non-users are holding out for more safety, education, and more integration in their lives before they experiment with generative AI:

70% of non-users would use generative AI more if they knew more about the technology.
64% would use generative AI more if it was more safe/secure.
45% would use generative AI more if it was integrated into the technology they already use.

Desk worker perceptions of generative AI

Workers are of two minds when it comes to generative AI, Salesforce research reveals. While they see the benefits of generative AI, they are also concerned about navigating the risks and learning the skills needed to take full advantage of the quickly evolving technology.

Three out of five workers (61%) currently use or plan to use generative AI.
More than two out of three (68%) say generative AI will help them better serve their customers.
Two out of three (67%) say generative AI will help them get more out of other technology investments, like other AI and machine-learning models.

However, many workers don’t yet know how to use generative AI responsibly. Half of the workers surveyed (54%) said they worry generative AI outputs are inaccurate, and 59% worry the outputs are biased.

Nearly three-quarters of respondents (73%) believe generative AI introduces new security risks.
Of those who said they plan to use generative AI, almost 60% said they don’t know how to do so using trusted data sources or while ensuring sensitive data is secure.

Business leaders also appeared to overestimate their ability to use generative AI responsibly. Salesforce’s research showed most C-suite leaders (83%) claim they know how to use generative AI while keeping data secure — compared to only 29% of individual contributors.

Clearly, companies have a gap to close when it comes to ensuring generative AI is effectively adopted and used. Fortunately, the data offers some insight. When asked how to establish generative AI as a trusted technology, employees identified four key elements to successfully using it in their role:

Human oversight (60%)
Enhanced security measures (59%)
Trusted customer data (58%)
Ethical use guidelines (58%)

IT professionals see vast opportunity in generative AI

Salesforce’s latest State of IT report found that 86% of IT leaders expect generative AI to soon play a prominent role at their organizations. And, the company’s March survey of 500 IT decision makers revealed that most (57%) believe generative AI is a “game changer.”

67% of IT leaders surveyed said they have prioritized generative AI for their business within the next 18 months.
33% said it was a top priority.

Alongside the enthusiasm, the March survey respondents were nearly unanimous (99%) in saying their business must take measures to use generative AI responsibly, and most senior IT leaders said their organization is not ready to deploy it.

Nearly two in three respondents (65%) said they can’t justify implementing generative AI yet. IT decision-makers cited several barriers to using generative AI, including:

Generative AI will introduce new security threats to their data (71%).
Their employees lack the skills to use it successfully (66%).
Generative AI can’t yet integrate with their business’s tech stack (60%).
Their organization lacks a unified data strategy (59%).

Marketers believe generative AI will transform their role but worry about accuracy

Despite how generative AI’s relatively new status as a mainstream technology, half (51%) of the over 1,000 marketers Salesforce recently surveyed are already using it or are experimenting with it at work. Another 22% of respondents said they plan to use generative AI soon – which, if that pans out, would mean nearly three out of four marketers surveyed could be using it today.

The most common use of generative AI among marketers is:

Basic content creation (76%)
Writing copy (76%)
Inspiring their creative thinking (71%)
Analyzing market data (63%)
Generating image assets (62%)

Marketers also believe generative AI will transform their role, freeing them up to focus more on strategic work:

Seven in 10 marketers (71%) expect generative AI will help eliminate busy work and allow them to focus more on strategic work.
They also predict generative AI will save them five hours of work per week – which amounts to over a month per year.*

Not without their reservations, marketers ranked accuracy and quality as the top concern among marketers (31%), followed by trust (20%), skills (19%), and job safety (18%).

Salesforce’s research also shows that many marketers need training to make the most of generative AI.

39% of marketers don’t know how to use generative AI safely.
43% said they don’t know how to get the most value out of it.
54% believe generative AI training programs are important for successfully using it in their role.
Despite that, 7 in 10 marketers said their employer does not yet provide generative AI training.

Image Modal

Generative ai snapshot series.

Uncover generative AI insights from over 4,000 full-time workers across the US, UK, Australia and India on Tableau.

Salespeople say generative AI will help sell more efficiently

While only about one-third of salespeople surveyed said they used or planned to use generative AI – compared to 51% of marketers – 61% of sales pros said they believe generative AI will help them better serve their customers. The same percentage of sellers also said they believe generative AI will help them sell more efficiently.

Salesforce research found that the skills gap may be affecting sales reps in two ways: It could be holding them back from adoption and instilling fears over job security.

Most sellers, 53%, said they do not know how to get the most value from generative AI at work.
Just under half said they do not know how to safely or effectively use generative AI – 49% and 47%, respectively.
Almost two out of five sales professionals (39%) worry they will lose their job if they don’t learn how to use generative AI at work.

Despite salespeople’s relative hesitance to use generative AI, the technology has already proven successful for early adopters. Of the respondents currently using it, 84% said it helped increase sales at their organization by enhancing and speeding up customer interactions.

Basic content creation (82%)
Analyzing market data (74%)
Automating personalized sales communications (71%)
Helping to generate sales reports (51%)
basic content creation (48%)
Analyzing market data (47%)

Service professionals think generative AI will help them serve customers faster

Among the workers surveyed in Salesforce’s research series, service professionals were the least likely to be using generative AI.

Just 24% of customer service said they were using generative AI for work.
Only 15% said they plan to use generative AI in the future.

Like sales pros, the skills gap may be behind service workers’ low adoption rate.

Three out of five service workers (60%) said they do not know how to get the most value out of generative AI at work.
Just over half of respondents said they do not know how to use the technology effectively or safely at work, 55% and 54%, respectively.

Despite the hesitance to use generative AI, nearly half of the surveyed service professionals (48%) worry they will lose their job if they don’t learn how to use the technology. However, of the service professionals currently using generative AI, 9 out of 10 say it helps them serve customers faster.

Service workers using generative AI were remarkably consistent in how they are applying the technology. Salesforce’s research found the leading generative AI uses among service professionals are:

Basic content creation (68%)
Create and personalize service communications with customers (68%)
Better automate customer service communications (67%)

They were similarly consistent in how they think generative AI will transform customer roles:

48% of service professionals said generative AI will improve customer self-service options.
47% said it will help automate customer service communications better.
46% said it would help analyze service data.

Key takeaways

Though each industry has unique use cases and perspectives around generative AI, several throughlines have emerged across Salesforce’s generative AI research:

Workers largely see the potential benefits of generative AI, with the potential for saving time spent on busy work to driving better customer experiences and increasing revenue.
Data security, accuracy, and trust in generative AI are points of concern among workers and needed to use the technology at work successfully.
Most respondents say their employers are not offering the training they need to adopt generative AI.

“We ask ourselves daily, what can our customers solve with AI?” said Clara Shih , CEO of Salesforce AI. “ For example, in marketing, it’s using AI to identify audience segments and auto-generate landing page images and copy. In sales, it’s summarizing key account changes and auto-generating prospecting emails that compel customers to open and respond. In customer support, it’s auto-generating case summaries and knowledge articles following a successful issue resolution. And for developers, it’s auto-generating test cases, comments, and even the code itself.”

She continued, “The business value comes from driving efficiency as well as consistently higher quality by augmenting every employee at every turn.”

For a full look at Salesforce’s industry research, visit the Stat Library
Find out how generative AI can help small businesses sell smarter
Learn the ABCs of AI with the generative AI glossary for business leaders here
See what generative AI means for CRM

*(5 hours per week on average saved using ‘generative AI’ x 52 weeks in one year) / 8-hour working day = 32.5 days or over a month per year saved on average

Just for you.

Now Available: Tableau Pulse Natively Embeds AI-Powered Insights into Salesforce, Helping Teams Make Data-Driven Decisions

Electrifying the Future: How Automakers and Energy Providers Can Use Salesforce to Supercharge the EV Transition

Unify your data, AI, and CRM with the Einstein 1 Platform

Explore related content by topic

Artificial Intelligence

Salesforce’s Four Keys to Enterprise AI Success

Salesforce Analysis Finds Businesses Are Excitedly Pursuing AI – But Worry About Trust, Data, and Ethics

How Is Gen AI Changing the World of Data Visualization?

Salesforce Report: Sales Teams Using AI 1.3x More Likely to See Revenue Increase

Get the latest Salesforce News

América Latina (Español)
Brasil (Português)
Canada (English)
Canada (Français)
United States (English)

Europe, Middle East, and Africa

España (Español)
Deutschland (Deutsch)
France (Français)
Italia (Italiano)
Nederland (Nederlands)
Sverige (Svenska)
United Kingdom (English)
All other countries (English)

Asia Pacific

Australia (English)
India (English)
Malaysia (English)
ประเทศไทย (ไทย)

© Copyright 2024 Salesforce, inc. All rights reserved. Various trademarks held by their respective owners. Salesforce, inc. Salesforce Tower, 415 Mission Street, 3rd Floor, San Francisco, CA 94105, United States

Get the latest Salesforce news

By subscribing, you confirm that you agree to the processing of your personal data by Salesforce as described in the Privacy Statement

Salesforce Research

Conversational AI Programming with CodeGen: Let AI Write Code For You

Links: Research Paper , Github

Can you imagine a machine writing an app for you, just by telling it what you want?

As futuristic as this scenario sounds, it’s actually here today.

Salesforce AI Research outlines conversational AI programming as a new paradigm that’s making this vision a reality, thanks to an AI system that writes software with you, in a conversation.

Introducing CodeGen: Turning Prompts Into Programs

The first step towards this vision is now here in the form of our large-scale language model, CodeGen , which turns simple English prompts into executable code. You don’t write any code yourself; instead, you describe what the code should do, in natural language -- and the machine writes it for you.

For a quick look at how it works, let’s ask CodeGen to solve the two-sum problem: find two integers in a list that add up to a certain number.

To begin, we simply prompt the model, in plain English, to solve the two-sum problem. As you can see in the brief video below, our CodeGen model generates functioning code that solves the problem correctly.

Some Quick Background: Terms, Definitions, Concepts

Before proceeding further, let’s define some of the terms and ideas used in this blog:

Programming or Coding: A multi-step process designed to get a machine to achieve a goal:

Translate a problem into a series of steps that solve it (the algorithm)
Translate that algorithm into a computer language (the program)
Run that program to see if it works (the test)
Find out which parts of the program did not work properly (debugging)
Revise the program (adjust for errors) and run it again (re-test)
Continue the run-debug-revise cycle until the program works (a working program runs successfully and solves the problem).

Conversational AI : Technologies enabling natural interactions between a human and a computer, via a conversation conducted in the human's native language.

Chatbots, voice assistants, and virtual agents are examples of conversational AI.

Types of Computer Programming

While software engineering concepts and methodologies have evolved over the past few decades (programming languages, web services, cloud computing, and so forth), the classical paradigm in which one writes the code (the underlying building block of software) has remained mostly untouched for decades. Since our research proposes a new way to do programming, it’s instructive to see how it compares to other means of doing programming:

In 1945, this is how the first programmable machine, the ENIAC, was programmed using plugboard wiring.
Today, this is how programs are written using formal languages with higher abstraction such as C, Python, or Java.
The classical, fundamental paradigm of specifying a problem in natural language and iteratively refining a solution in a formal or programming language until the specification is satisfied remains the predominant method of programming today.
Most of today’s popular computer languages are like this; coders write in a higher-level language, and a compiler generates low-level code; this saves time and effort, since we humans don't have to worry about all the low-level details.
Interactive programming : Coding a program (or parts of a program) on-the-fly, while that program is running.
Since it combines conversational AI (interactive human-to-machine dialogue) and automatic programming (the system automatically creates the program based on a higher-level language: your conversation!), we call what CodeGen does conversational AI programming.

A Different Kind of Coding Problem: (Learning a New Language) = [Barrier]

Up till now, we have had two ways to get computers to do useful work:

use pre-existing computer programs that do what you want the machine to do
write a new program to do it.

Option 1 is great, when the computer programs you need are available.

But Option 2 has a built-in barrier : if the type of program you need does not exist, the task of creating new programs has always been limited to those who can speak the computer’s language; you must learn at least one programming language and apply that knowledge to write programs. In other words, to get new programs, you have to know how to translate what you want to do into computerese , so the computer will understand what you want it to do. This bottleneck applies not only to situations where you want to create a program for yourself , but also when you want to create programs for others - in coding jobs , for instance.

Here are three of the major limitations of the current programming paradigm:

Time-consuming: one must learn a programming language and apply the knowledge correctly
Difficult: some find the process of learning this new language to be an arduous task, and some cannot get through the training successfully
Expensive: have you seen the cost of coding schools?

These factors often hinder or discourage the education and development of new programmers, especially among people in historically disadvantaged groups. In other words, traditional programming often presents people with a different kind of "coding problem" -- not one given on a test, but rather a formidable real-world obstacle that many simply cannot solve.

However, the good news is that there is another way.

The CodeGen Solution: Make Coding as Easy as Talking

What if you could just tell a machine the kind of program you need - just use your native language to describe your needs to a computer, and it would generate the code that does what you wanted? That’s the amazing promise of conversational AI programming: CodeGen makes programming as easy as talking .

Here’s an analogy to help illustrate the concept. When you order dinner in a restaurant, instead of having to know the correct ingredients to make your desired dish, and then cooking it yourself, you just tell the server what you want, and they prepare and bring it to you. Say the dish you want in a short sentence, and you get it without any involvement from you in the creation of the meal - no need to specify any ingredients or cooking steps, and you don’t need to know any special culinary terms. The restaurant acts like an intelligent system, translating your plain-English request (order) into a sequence of steps that takes basic food ingredients and generates the outcome (cooked dish) you asked for. Now imagine you’re “ordering” computer code instead of a meal, and you have the basic idea behind CodeGen.

Our implementation of conversational AI programming provides a glimpse into the future of democratizing software engineering for the masses. An “AI assistant” translates English descriptions into functional and executable Python code - allowing anyone to write code, even if one knows nothing about programming. The underlying language model, CodeGen, enables this conversational paradigm and will be made available as open source to accelerate research.

The Full Vision: Interactive Conversation with Computer Creates Code

This new paradigm in programming takes the form of a simple yet highly intelligent dialogue. In the concept’s full implementation (our vision of how it would work in its ultimate form), a typical fully-interactive conversation about your desired code would flow as follows:

The Current State: Conversational AI Programming with CodeGen

While the above (fictional) conversation example helps illustrate the full conceptual vision, let’s turn to some real-world examples - the concrete realization of the concept as it exists today in CodeGen. Let's start by revisiting the two-sum problem we introduced at the start of this blog:

Note that this time, we don’t stop once CodeGen generates working code to solve the problem - we ask the model to try again , and solve the problem using a hash map . This example illustrates some of the groundbreaking capabilities of our system: we can continue our conversation, refer back to “the problem” (a backreference , which CodeGen understands), and request that the model try a new approach (the hash map), in the hopes of getting an even better solution. (In our restaurant analogy, this would be like giving additional instructions to the server about your order, like “use egg whites only” or “use margarine instead of butter.”)

And it works: CodeGen succeeds in generating new code that uses a hash map, and in so doing, the new solution runs in linear time - O(n) - much faster than the original solution, which was O(n**2).

The Two Sides of CodeGen: For Non-Coders and Programmers Alike

The above “hash map” example illustrates a key aspect of CodeGen: while anyone can use CodeGen to build software from scratch, even non-coders, it does help to have some programming knowledge in certain cases. For example, knowing coding concepts can help you think of followup commands to give CodeGen, suggesting new avenues to explore while building the code (like using hash maps, or recursion - or not using these techniques).

While the vision is to create optimal programs for any problem by just telling the machine what you want, without needing any coding knowledge, the reality is that some programming knowledge can often help, in order to guide CodeGen to a good solution. This is especially true for more complex problems, where having the user suggest different approaches may help the software find a working solution - or a more efficient one.

Still, even for experienced coders, CodeGen makes getting to a functioning solution faster and easier, and allows rapid exploration of alternate methods. In other words, CodeGen is beneficial for all levels of programmers.

The Details: An In-Depth Look at How CodeGen Works

Approach . Salesforce AI Research trained CodeGen, a 16-billion parameter auto-regressive language model, on a large corpus of natural and programming languages. Two aspects are of particular interest: (1) sampling executable code by scaling the size of the model and dataset, (2) emergence of conversational capabilities.

Scaling . The large size of this model is motivated by the empirical observation that scaling the number of model parameters proportional to the number of training samples appears to strictly improve the performance of the model. The phenomenon is known as the scaling law . We leverage this law to learn a model which can translate a natural language (English) to a programming language (Code) with high accuracy. That is, the model is capable of not only generating reasonable code, but also executable code; the generated code is of such high quality that it can be immediately executed without revisions by a programmer, which allows even a non-professional audience to “write” code.

Conversation . Having a conversation appears a rather trivial task for humans. We implicitly keep track (or a memory) of the past conversation, resolve references to previously mentioned elements, and incrementally build a mental picture or story of the discourse. For machines, holding a realistic conversation is one of the grand challenges of our time. Testing if a machine possesses human capabilities or can fool a human into believing she or he is holding a conversation with another human being is known as the Turing Test. While in the first iteration of our research, the model replies in a formal language (i.e., the programming language) and not a natural language, later incarnations will be in the form of a multi-turn discourse in natural language, so that the model may resolve ambiguities as in “May I solve this problem with algorithm A, B, or C?”. Surprisingly, modeling such conversation in conjunction with the scaling laws turned out to be rather simple, where simplicity is a rather desirable property (see Rich Sutton’s “The Bitter Lesson”).

Specifically, a conversation of several consecutive questions (by the human in natural language) and answer (by the machine in programming language) is concatenated into a single long sequence. Based on this context of the past conversation, an auto-regressive decoder model samples the next response conditional on the past pairs of questions and answers. The fact that conversational capabilities emerge with such a naively simple approach (given sufficient data and model size) was surprising.

Picture the example shown earlier, in which first a problem is stated and subsequently the question (or specification) is refined:

“Solve the two sum problem”

“Solve the problem using a hash map”

While solving the first request can be understood as a form of pseudo-retrieval of examples in the observed training data (think of a database query), the second request involves resolving the backreference of “the problem” to “the two sum problem”, and requires a shallow form of understanding the previously generated code to rewrite it using a hashmap. This phenomenon is crucial as the underlying model was never specifically trained to hold a conversation or revise code. These conversational and problem solving capabilities “emerged” naturally.

Societal Benefits and Impact: Why Conversational AI Programming is Important

Benefits for Next-Gen Software Development: Programs of the Future Need This

While programming is a useful skill today, in the next decade programming will be a necessity in many tech jobs, including at Salesforce. The world needs more and more code, in every aspect of society, and these programs are getting increasingly complex. Hence, systems like CodeGen (which help speed up the programming process while making it easier and more manageable) should play an integral role in completing increasingly large coding projects, as well as bringing a whole new generation of programmers into the world of coding to achieve these goals.

But there is another issue on the horizon: what happens when future programming needs become so complex that the skills needed to create these programs outstrip human capabilities ? Digital ecosystems are evolving into systems with ever-increasing functional complexity, and at some point these systems’ complexity may increase beyond our capacity to understand them, let alone build them. We may soon get to the point where projects require technology such as conversational AI programming, in order to create the mega-complex software systems of the future -- both on the massive scale that will be required, and in a timeframe that would be impossible for a team of human programmers to produce on their own.

In short, rapidly increasing code complexity requires a new paradigm. Next-gen programming needs, both at Salesforce and at other organizations, seem destined to make conversational AI programming systems like CodeGen essential to our future.

Benefits for Society: CodeGen Democratizes Programming

A major part of Salesforce’s mission is to develop technology that can help all of society, not just the company , and that is exactly what this research does. Many groups will benefit from the conversational AI programming revolution that CodeGen represents. Here are some examples.

Enhancing equality and equity . Opening up coding to all - democratizing access to the world of creating programs - will help bring traditionally disadvantaged groups into the world of programming, leading to increased career opportunities and incomes for such groups.

Education/teaching/learning . Kids will learn to program interactively with “AI teachers” as their companions, create worlds and games in a discourse in their native language, while learning and absorbing how to translate their ideas into programming languages.

Software professionals: engineers, data scientists, developers . Software engineers will understand the architecture, design patterns, and summarize critical paths of legacy systems with the aid of “AI assistants”. Analysis of complexity in space and time, security vulnerabilities, design patterns, refactorings, or test generations is supported by an artificial pair-programmer.

Non-software professionals . Business analysts will integrate complex external data sources and systems, correlate and normalize data, perform exploratory analysis, and visualize findings in conjunction with “AI analysts”.

In general, the democratization of coding should reap society-wide rewards .

The new paradigm of conversational AI programming will lead to a disruptive transformation in software engineering. Today, computer science education requirements act as a barrier to the world of programming, but tomorrow the floodgates will be opened for everyone to transform an idea into code - no programming knowledge required.
We expect this disruption to be on the scale of other tech-based breakthroughs such as autonomous vehicles, or even the printing press.
Just as the printing press revolutionized the world by bringing publishing and reading books to the masses, so too will creating programs using natural language revolutionize the world, by letting anyone make programs that will benefit themselves, or others - and do so at much faster speed.
Democratization of knowledge, speeding up creation of new knowledge
Democratization of programming, speeding up creation of new programs.

The Bottom Line

The paradigm of conversational AI programming (coding by talking) turns the notion of writing code for a machine on its head. Rather than requiring a human to write code for a machine, the machine generates code for the human (automatic programming) via a discourse between human and machine (conversational AI). The human merely needs to explain the desired functionality in plain English, while the machine writes the code based on that description.
Most users can solve relatively simple coding problems with little or no prior programming knowledge.
More complex cases may require some knowledge of programming or basic computer science concepts, in order to help guide the system as it searches for a solution (i.e., working code that solves the stated problem).
Still, even for experienced coders, CodeGen makes getting to a functioning solution faster and easier, and allows rapid exploration of alternate methods.
This new approach democratizes access to the world of writing software, allowing anyone to develop apps in conjunction with an “AI assistant” or “teacher” without the need to learn programming in the traditional way.
Opening up coding to all will help bring traditionally disadvantaged groups into the programming world, leading to more career opportunities and higher incomes for such groups.
Our research is one of the first steps towards implementing the wider vision: enabling machines to leverage in-depth natural language conversations with humans to write code - and do it faster, with fewer errors, in a manner that makes it easy for all. There will be challenges along the way, but these efforts will yield a wide range of new applications as we pursue this exciting research direction.
We will open source the code. Part of the mission of Salesforce is to create and publish important research that others may benefit from, so our work on conversational AI programming will be available as open source code.

Explore More

Salesforce AI Research invites you to dive deeper into the concepts discussed in this blog post (links below). Connect with us on social media and our website to get regular updates on this and other research projects.

Deep Dive: Read more about CodeGen in our research paper
Code : Check out the CodeGen Github page
Feedback ? Questions ? Email Erik Nijkamp at [email protected]
Follow us on Twitter: @SFResearch - and read our CodeGen tweet .
Learn more about all of the projects we’re working on at our main site : https://www.SalesforceAIResearch.com

About the Authors

Erik Nijkamp is a Research Scientist at Salesforce AI Research. His research emphasis is on large-scale generative models and representation learning with applications in NLP and computer vision. Prior to Salesforce, he was a PhD student under Prof. Song-Chun Zhu and Prof. Ying Nian Wu at UCLA.

Donald Rose is a Technical Writer at Salesforce AI Research. He works on writing and editing blog posts, video scripts, media/PR material, and other content, as well as helping researchers transform their work into publications geared towards a wider (less technical) audience.

3 Ways Salesforce Takes AI Research to the Next Level

Written by Shelby Heinecke and Scott Nyberg .

In our “Engineering Energizers” Q&A series, we examine the life experiences and career paths that have shaped Salesforce engineering leaders. Meet Shelby Heinecke, a research manager for the Salesforce AI team. Shelby leads her diverse team on a variety of projects, ranging from identity resolution to recommendation systems to conversational AI, supporting cutting-edge initiatives spanning academia and Salesforce products.

When did you know you were interested in pursuing a career in artificial intelligence?

While pursuing my math PhD, I had several summer internships, ranging from traditional software engineering to machine learning. These inspired me to pivot my PhD research from traditional math topics to computer science-inspired topics. That ultimately led me to machine learning and paved my path to Salesforce.

Machine learning and artificial intelligence (AI) are a good mix of my personal interests. Both incorporate interesting math and engineering problems that can have immediate real-world impact.

Shelby explains what drove her decision to join Salesforce.

What is your AI team’s role and mission here at Salesforce?

My team’s role and mission focuses on pioneering new AI approaches for personalization. That is a big part of what we do.

To provide some context, when you use apps and services, you expect a certain level of personalization. The experience is similar when you are on social media — you expect to see articles and media recommended to you according to your preferences. That is an example of personalization.

From a business standpoint, personalization improves user experiences and engagement. And of course, personalization is not limited to social media — any data-based services you interact with can be personalized, or tailored, to your individual needs.

What are some interesting AI projects your team is working on?

We currently focus on three key projects: Identity resolution, recommendation systems, and conversational AI.

With identity resolution, the first step in personalizing any service involves collecting and preparing data; this often means merging and aligning multiple diverse data sources. We build state-of-the-art AI solutions to intelligently execute that step.

A look at some of the ways in which individuals may appear multiple times across data sources.

Recommendation systems are in nearly every website you use today — it is the engine that follows your feeds to optimize recommendations for media, who to follow, products, ads, and articles. These recommendations are personalized according to your previous clicks, views, purchases, etc. My team develops state-of-the-art recommendation models for academic research and Salesforce products.

Lastly, we develop cutting-edge conversational AI approaches for conversational systems. You have seen these systems emerging everywhere — think customer service bots on websites, etc. By leveraging the power of personalization, companies can learn about their customers from their conversation history, thus making these conversational systems more effective.

How has your leadership style evolved since joining Salesforce?

I first joined Salesforce as a research scientist, which is tantamount to being an individual contributor. In that role, I grew as a technical leader on projects ranging from identity resolution to recommendation systems. It was an awesome experience. Recently, I became a manager, and I am also growing as a people leader.

How would you describe the culture within your AI team?

As a research lab, the foundation of our culture is innovation. From developing leading edge models for Salesforce products to pushing the boundaries of what’s been achieved in academic research, we live and breath at the frontier of AI. In fact, we are often defining new frontiers.

Beyond being innovative, what also makes us effective as a lab is that we are collaborative and agile. We work closely with a range of Salesforce technical and non-technical teams — rapidly developing new solutions in face of any real-world constraints.

Shelby shares a recent engineering challenge that kept her up at night.

How important is diversity to you and how does it strengthen your AI team?

Diversity is a critical ingredient for innovation and problem solving. The coolest ideas come from new perspectives gained from our individual, unique experiences. Most importantly, for the complex problems that our team faces in AI, new perspectives and new ideas are necessary.

My team consists of researchers with diverse expertise, ranging from natural language processing (NLP) to recommendation systems to machine learning theory. Collectively, our research spans empirical to theoretical, and our experiences everything from academia to industry.

With our unique perspectives, we come together to solve challenging AI problems.

What is your favorite sport to play?

I love CrossFit. I discovered it around a year and half ago. I love that it is infinitely challenging and community-oriented. For most workouts, you are technically performing individually, but there is so much fun and community in pushing your limits with others beside you who strive to do the same. We push each other to be the best we can be.

It is interesting how some of CrossFit’s themes overlap with my Salesforce AI team. As individual researchers, we want to be the best we can be. So I think we try to push each other to push our boundaries — enabling everyone to maximize their fullest potential.

Hungry for more AI? Read this blog to learn how Salesforce developed an AI-based identity resolution system, enabling customers to precisely identify their customers across highly diverse data sets.
Check out our Technology and Product teams to learn how you can get involved.
Stay connected – join our Talent Community !