+

SURVEY REPORT

AI, Today in
Manufacturing &
Energy
Introduction

In June 2023, Dataiku and Databricks surveyed 400

senior AI professionals in large companies around
the world.1 This research aimed to measure the
excitement around this new era of technology and
better understand the larger environment in which we
find ourselves today, from tech stacks to spend on AI
tools and services, AI use case reach, and more. We’ve
compiled some of the top statistics of the state of AI in
the manufacturing industry.

1 Respondents are manager level or higher, come from the U.S., EMEA,
APAC, and work at organizations with >$3 billion in annual revenue. See p. 15
“About the Survey” for additional details about respondents and methodology.

SURVEY REPORT 2 DATAIKU + DATABRICKS

 AI Maturity in Manufacturing vs. Across All Industries

Scaling AI Maturity in
13% 18%

27%

Manufacturing 30%

35%
31%

24% 20%

62%
1% 1%
Manufacturing, Industrials, Oil & Gas Total

of manufacturing organizations are None: We don't leverage advanced analytics (data science or Al).
either establishing or exploring/ Exploring / Experimenting: We're still experimenting with our first projects and use cases.
experimenting when it comes to AI
Establishing: We've proved value with several use cases and we're working on building processes to scale out our
adoption, marking them closely in line
ability to execute.
with the all-industry average (61%) for
that category. Expanding: We're working on expanding the use of data science and Al beyond a central team and into different
functions and departments across the organization.

Embedding: We already have people across different functions and departments using data science/Al, and we're
working on embedding into even more activities, making data part of the DNA of the organization.

Q7: What best describes the level of adoption of data science and/or AI at your organization?

SURVEY REPORT 3 DATAIKU + DATABRICKS

 Top AI Techniques in Manufacturing

Experimenting I don't know what this is Not using and don't plan on using

Experimenting Actively using or have used

Not using, but plan to in the next year

With Some (But Predictive analytics 4% 6% 27% 63%

Not All) New AI Forecasting 5% 8% 33% 53%

1%
Technologies Image recognition 17% 23% 23% 36%

1%
Natural language processing (NLP) 6% 19% 18% 29% 28%

Fraud/anomaly detection 22% 22% 26% 28%

2%
Large language models (LLMs), Self-supervised learning 3% 25% 25% 25% 22%

reinforcement learning, and Segmentation 7% 19% 18% 35% 21%

segmentation are the most up-and- Product recommendation 26% 18% 34% 21%

1%
coming techniques. Edge Al / tinyML 9% 25% 21% 25% 20%

Foundation models 10% 22% 20% 29% 19%

Many manufacturing firms aren't — and Reinforcement learning 4% 19% 24% 37% 16%
don’t plan on — using self-supervised Transfer learning 9% 22% 27% 26% 16%
learning, Edge AI / tiny ML, product Federated machine learning 13% 22% 22% 28% 14%
recommendation, voice recognition,
Voice recognition 30% 28% 28% 14%
and feature store. This is the only
Casual Al 9% 16% 31% 31% 13%
industry with such high figures of
Generative Al / large language models (LLMs) 4% 19% 21% 44% 12%
people not planning to use some of
Synthetic data 9% 18% 31% 30% 12%
today's hottest AI techniques.
Feature store 15% 34% 17% 25% 9%

Q15: What data science, machine learning, or AI techniques are you using for any use case in your organization?

SURVEY REPORT 4 DATAIKU + DATABRICKS

 Manufacturing Firms
Plans to Use Generative AI or Large Language Model (LLM) Technology Over the Next Year Show Caution When it
in Manufacturing vs. Across All Industries
Very Likely Likely Unsure Unlikely Very Unlikely
Comes to AI

52%
18%
25%

34%
of manufacturing firms plan to use Generative AI
39%
or LLM technology over the next year (marking
them below the average across all industries of
64%).
31%
23%

14%
14%
11%
3% 2%

Manufacturing, Industrials, Oil & Gas Total

of manufacturing firms are unlikely to use

Q20: Are you planning to use Generative AI or large language model (LLM) technology for your business in the coming year? Generative AI or LLM technology over the next
year.

SURVEY REPORT 5 DATAIKU + DATABRICKS

 AI Use Cases in Manufacturing

Lines of Business Not Applicable Not planning to develop these use cases in the next year.

With Top AI Use cases in this area are on our road map for the next year. Developed or are currently developing use cases in this area.

Use Cases Operations 10% 31% 59%

Opportunities Information Technology (IT) 10% 32% 57%

1%
Research and Development (R&D) 7% 15% 34% 44%

Marketing & Sales 19% 39% 41%

1%
The following are up-and-coming
Customer Service 4% 19% 41% 36%
areas for use cases for AI in the
Procurement 5% 28% 34% 33%
manufacturing industry:
Accounting & Finance 3% 28% 38% 31%

Human Resources (HR) 4% 38% 37% 21%

Strategic management
Strategic Management 8% 28% 44% 20%
Customer service
Legal & Compliance 13% 52% 29% 6%
Procurement
Q14: For what lines of business has your organization developed advances analytics, data science, and/or AI use cases?

SURVEY REPORT 6 DATAIKU + DATABRICKS

Positive Attitudes Towards ChatGPT

61%
Despite their cautious attitudes towards Generative AI, manufacturing firms appear more optimistic regarding
ChatGPT.

Attitudes About ChatGPT in Manufacturing vs. Across All Industries

of manufacturing organizations believe
ChatGPT is a revolution or will have many
It's a revolution and will change the way we work.
valuable applications and that companies
It will have many valuable applications and companies should start developing those now. should start developing them now.
It could have valuable applications, but its not clear what those are.

It's a distraction from more useful applications of Al.

Other

30% 24%

31% 37% Only 4%

believe ChatGPT to be a distraction from
34% 33% more useful applications of AI.

1% 4% 5%
Manufacturing, Industrials, Oil & Gas Total
Q23: Which of the following most closely aligns with your personal feelings about ChatGPT and similar technology?

SURVEY REPORT 7 DATAIKU + DATABRICKS

Manufacturing Has ROI on AI per $1 Spent in Manufacturing
Room to Grow on ROI Cannot Discuss >$10

17% 2%

38%
$5-10
<$1
14%

13%

of these firms expect a return of $2 to $5 for

each $1 spent (compared to 33% across all $1
industries).
1%
$2-5
$1-2
38%

16%
15%

of manufacturing firms expect an estimated

return of $5 to >$10 for each $1 spent on Q10: Approximately what return do you deliver for each $1 spent on data, analytics, and AI initiatives?

AI initiatives (compared to 21% across all

industries).

SURVEY REPORT 8 DATAIKU + DATABRICKS

 Data and AI Technology Tools Budget in Manufacturing vs. Across
All Industries – Next 12 Months

Budgeting Less Than the Average Don't know < $1 million $1 - $5 million $5 - $20 million > $20 million

Across All Industries for AI Efforts Manufacturing,

Industrials, Oil & 14% 14% 34% 21% 17%
Gas

Total 11% 9% 33% 30% 17%

38%
of manufacturing firms are budgeting high amounts for AI tools. That's
Q16: Approximately how much is your organization budgeting to spend on data and AI technology and tools in the next 12 months?

between $5 and >$20 million on data and AI technology and tools in the next
12 months. Data and AI Services Budget in Manufacturing vs. Across
All Industries – Next 12 Months

Don't know < $1 million $1 - $10 million $10 - $50 million > $50 million

43% Manufacturing,
Industrials, Oil &
Gas
14% 16% 43% 19% 8%

of these firms are budgeting high amounts for AI services. That's between $1
and $10 million on their data and AI services in the next 12 months.
Total 11% 14% 43% 22% 10%

Q17: How much is your organization budgeting to spend on data and AI services (consultants, system integrators, etc.) in the next 12 months?

SURVEY REPORT 9 DATAIKU + DATABRICKS

 Consolidating Data, Analytics, and AI Tech
Stacks
The tools most adopted in the industry are:

BI and data visualization tools (85%) Data warehouses (82%) Data lakes (70%)

Technologies Part of Data, Analytics, and Al Stack in Manufacturing vs. Across All Industries
% selected “Yes, this is currently part of our data, analytics, & AI technology stack.”

20% 40% 60% 80% 100%

BI or data visualization tool(s) 85% 84%

Data warehouse 82% 82%
Data lake 70% 72%
Data catalog 52% 59%
Notebooks or code-based data science platform 35% 49%
End-to-end, analytics and Al platform (e.g., Dataiku) 36% 44%
A dedicated MLOps platform 22% 35%
An AutoML solution 18% 24%

Manufacturing, Industrials, Oil & Gas Total

Q18: Which of the following are a part of your data, analytics, and AI technology stack?

SURVEY REPORT 10 DATAIKU + DATABRICKS

Confrontations With
Data Quality and
Operationalization Barriers Preventing Organizations from Delivering More Value from Data, Analytics,
and AI in Manufacturing

The top four barriers to delivering more value from Rank 8 Rank 7 Rank 6 Rank 5 Rank 4

data, analytics, and AI in the manufacturing industry Rank 3 Rank 2 Rank 1 - Biggest/most present barrier

are as follows:
1. Data quality and access Lack of quality data or ability to easily
7% 9% 11% 15% 17% 23%

2%
2. Project operationalization and iteration access the nght data.

3. Lack of clear business cases Inability to quickly operationalize and 4% 7% 5% 11% 16% 19% 19%
iterate on data. analytics. and Al projects.
4. Too much complex infrastructure Lack of data talent or data-literate
workforce.
6% 12% 9% 7% 9% 23% 13%

Infrastructure that is too complex. 7% 12% 12% 6% 17% 9% 14%

Manufacturing organizations share the same Lack of visibility and control of data,
analytics. and Al projects across the 3% 10% 14% 19% 17% 7%

1%

2%
top barriers as the average across all industries: organization.

Cost. 18% 11% 11% 9% 6% 4% 8%

1%
• 53%
53% of them ranked data quality and Lack of a clear business case. 12% 6% 5% 12% 9% 8% 15%
access as as one of their top three barriers
Q11: Please rank all barriers preventing you from delivering more value from data, analytics, and AI, where 1 is the biggest/most present barrier.

• 56%
56% of total respondents ranked project
operationalization and iteration as one of
their top three barriers

SURVEY REPORT 11 DATAIKU + DATABRICKS

 Having a C-Suite team that includes at least 1 data Providing company-wide education on data-related
leader topics

Below Optimum 47%

52%

Data Practices 60% 57%

and Culture of
the Industry
Having the right level of visibility into all AI projects
The manufacturing industry is well Having clear owners responsible for data quality
across the organization
below average across all industries in
the following efforts:
39%
50%
Manufacturing,
Industrials, Oil & Gas 43%
55%
Average Across All Industries

SURVEY REPORT 12 DATAIKU + DATABRICKS

Majority Calls For More
AI Regulation

52%
of manufacturing firms believe AI requires
more official regulation (via laws, government
oversight, etc.).

SURVEY REPORT 13 DATAIKU + DATABRICKS

Low Leadership Awareness of AI
Risks and Benefits in Manufacturing Belief That Leaders Across All Industries Understand the
Risks and Benefits of AI

Yes Unsure No

58% 37% 42%

54%
61%
of manufacturing organizations believe fears regarding AI are somewhat or
totally justified (just above the average across all industries of 55%). 16%
17%
16%
15%
47% 41%
30%
Less than half (42%) of manufacturing firm respondents think the leaders at their 24%

organizations (c-suite — CEO, CMO, CFO, etc.) understand the risks and benefits of AI.
Exploring / Establishing Expanding Embedding
Experimenting
In terms of AI adoption, over half of the most mature organizations (expanding or
Q22: Do you think the leaders at your organization (c-suite — CEO, CMO, CFO, etc.) understand the risks and benefits of AI?
embedding) across all industries believe that their leaders understand AI risks and
benefits.

SURVEY REPORT 14 DATAIKU + DATABRICKS

About the Survey

BACKGROUND & OBJECTIVES

Dataiku and Databricks conducted this joint survey to:

• Learn about the differing levels of adoption, sentiment, and barriers to AI in general.
• Understand the perception and level of adoption specifically around new
technologies like Generative AI and Large Language Models (LLMs).
• Assess trends in tech stack and tooling in today’s largest organizations.

SURVEY REPORT 15 DATAIKU + DATABRICKS

SAMPLE COMPOSITION

Length of Interview (LOI): ~8 minutes

Fielding: May 30, 2023 - June 8, 2023

• Data has been reported among total respondents unless otherwise noted.
• Small base sizes (n<30) are noted with an asterisk (*).

Decision-making authority when it comes to selecting and

Approximate annual revenue of your organization in 2022 purchasing data, analytics, business intelligence, and/or AI
software platforms.

SURVEY REPORT 16 DATAIKU + DATABRICKS

 +

Dataiku Is the Platform for Everyday AI.

With Dataiku, data experts and domain experts design, develop, and deploy new AI capabilities at
all scales and in all industries. Hundreds of companies worldwide use Dataiku to drive value from
data and AI with diverse use cases, from predictive maintenance and supply chain optimization
to quality control in precision engineering, marketing optimization, Generative AI use cases, and
everything in between. To learn more, follow Dataiku on LinkedIn.

Databricks Is the Data and AI Company.

More than 10,000 organizations worldwide — including Comcast, Condé Nast, and over 50%
of the Fortune 500 — rely on the Databricks Lakehouse Platform to unify their data, analytics
and AI. Databricks is headquartered in San Francisco, with offices around the globe. Founded
by the original creators of Apache Spark™, Delta Lake and MLflow, Databricks is on a mission
to help data teams solve the world’s toughest problems. To learn more, follow Databricks
on Twitter, LinkedIn and Facebook.

SURVEY REPORT 17 DATAIKU + DATABRICKS

©2023 dataiku | dataiku.com
 AI & Dataiku:
The Future of Manufacturing
From production quality control to electricity forecasting and more, Dataiku helps you pave the way
for Industry 4.0, elevating all parts of the business with AI (including Generative AI).

LET'S GO