Elastic (ESTC) CEO on How the Company Uses A.I.

Summary

In a recent interview, Ashutosh Kulkarni, CEO of Elastic Data Analytics, discusses the company's growth and innovative approach to AI and large language models (LLMs). He emphasizes the importance of maintaining privacy by keeping proprietary data secure while leveraging AI to extract insights from unstructured data. Kulkarni highlights the significance of their retrieval augmented generation model, which enhances the accuracy of AI-generated responses without exposing sensitive information. With a focus on organic growth and potential acquisitions, he shares optimism about Elastic's future in the AI landscape, underscoring their strategic plans for expansion and profitability.

Takeaways

• 🚀 Elastic Data Analytics is positioned as a search AI company, helping businesses extract insights from messy and unstructured data.
• 📈 The company reported a 20% growth in Q4 and a 32% growth in its cloud business, reflecting consistent performance.
• 🔍 Their platform allows businesses to use their private data securely, enhancing AI applications without compromising proprietary information.
• 🛡️ Retrieval augmented generation is a key mechanism that provides context to large language models while maintaining data privacy.
• 🤖 Large language models are trained on publicly available data, making it essential for businesses to protect their proprietary information.
• 👥 Elastic has a strong internal talent pool in machine learning and continues to hire experts in the field to enhance its capabilities.
• 💼 The company has over $1 billion in cash, which may be used for acquisitions or to support organic growth.
• 🌱 Elastic aims to build a generational company, focusing on long-term growth and profitability while capturing market share.
• 📊 Operating margins have improved, with non-GAAP operating margins projected to increase from 11% to over 12% in the current year.
• 🔮 The CEO emphasizes the massive opportunity within the AI space and the importance of leveraging business strengths for future success.

Q & A

What is the core business model of Elastic Data Analytics?

Elastic Data Analytics positions itself as a search AI company, focusing on helping businesses extract insights from unstructured and semi-structured data.

How does Elastic integrate AI into its search tools?

Elastic utilizes a mechanism called retrieval augmented generation to provide private data context to large language models, allowing businesses to build generative AI applications securely.

What recent growth metrics did Elastic report for Q4 of fiscal 2024?

In Q4 of fiscal 2024, Elastic reported a 20% growth for the quarter, with its cloud business experiencing a growth of 32%.

What concerns might clients have regarding data privacy when using AI models?

Clients are concerned about the potential risk of handing over proprietary data to large language models, which have only been trained on publicly available information.

How does Elastic ensure data security for its clients?

Elastic allows clients to keep their private data within their environment, ensuring that privacy controls are defined by the clients themselves.

What is the significance of the term 'retrieval augmented generation' in Elastic's services?

Retrieval augmented generation refers to Elastic's method of providing real-time context from private data to large language models, allowing for precise and relevant answers without compromising data privacy.

How does Elastic plan to build its workforce to support AI integration?

Elastic aims to grow its workforce by hiring new talent with machine learning expertise and leveraging existing talent within the company, along with considering technology tuck-in acquisitions.

What strategies does Elastic have for utilizing its significant cash reserves?

Elastic plans to use its cash reserves for both organic growth and potential acquisitions to capture more market share and support long-term growth.

What are the expected operating margins for Elastic in the current fiscal year?

Elastic anticipates its non-GAAP operating margin to exceed 12% in the current fiscal year.

Why is Elastic's search platform considered essential for businesses?

The search platform is crucial for businesses because it connects their proprietary data with large language models, enabling them to extract valuable insights without exposing sensitive information.

Outlines

00:00

🤖 Harnessing AI for Business Growth

In this segment, the discussion revolves around the integration of AI and Language Model Systems (LMS) at Elastic Data Analytics, led by CEO Ashutosh Kulkarni. The company has experienced a notable growth of 40% in the past year, showcasing its strength in AI-driven search solutions. Kulkarni describes Elastic as a search AI company, emphasizing its role in helping businesses, both large and small, derive insights from unstructured and semi-structured data. The company focuses on privacy and security by allowing clients to maintain control over their private data while leveraging generative AI applications. Kulkarni also highlights a 20% growth in Q4 and a 32% increase in the cloud business, asserting that their innovative approach positions them well in a rapidly evolving market.

05:01

🔍 The Future of Search and Data Privacy

This paragraph addresses the challenges and opportunities presented by the rapid development of language models and AI technologies. Kulkarni underscores the importance of proprietary data for businesses, arguing that many organizations are hesitant to share sensitive information with external language models. He explains that Elastic's platform enables real-time connections between private data and large language models, ensuring that responses are accurate and contextual without exposing proprietary information. This capability reduces the risk of data leakage while enhancing the precision of the responses generated, positioning Elastic's technology as essential for companies looking to harness AI responsibly.

Keywords

💡AI (Artificial Intelligence)

Artificial Intelligence (AI) refers to the simulation of human intelligence processes by machines, particularly computer systems. In the context of the video, AI is positioned as a transformative force in data analysis, enabling companies to extract meaningful insights from complex data sets. Ashutosh Kulkarni emphasizes that AI offers unprecedented ways to discover value from unstructured data, which is crucial for businesses aiming to maintain a competitive edge.

💡LMS (Large Language Models)

Large Language Models (LLMs) are AI models trained on vast amounts of text data to understand and generate human-like language. Kulkarni discusses how Elastic Data Analytics integrates LLMs into their search and analytics offerings, enabling businesses to create generative applications that leverage their private data without compromising security. This is illustrated when he explains the mechanism of retrieval-augmented generation, which enhances the functionality of LLMs by incorporating specific, proprietary information.

💡Data Privacy

Data Privacy refers to the proper handling, processing, storage, and usage of personal or proprietary data, ensuring that individuals' information is protected. In the video, Kulkarni stresses the importance of data privacy for businesses, noting that companies are hesitant to share their sensitive data with external LLMs. He highlights that Elastic allows organizations to maintain control over their data while still leveraging advanced AI capabilities, which aligns with the growing demand for privacy-first solutions.

💡Search AI

Search AI combines search technology with artificial intelligence to enhance the retrieval of information from complex datasets. Elastic identifies itself as a search AI company, providing tools that help organizations find insights from messy, unstructured data. Kulkarni explains how their platform allows businesses to apply AI to their internal data in a way that enhances decision-making and operational efficiency.

💡Retrieval-Augmented Generation

Retrieval-Augmented Generation is a mechanism that allows AI models to access and utilize specific documents or data during the generation of responses. Kulkarni describes how this approach enables their LLMs to provide accurate answers based on a company's proprietary information while avoiding the risk of hallucination or incorrect responses. This concept is central to Elastic's offering, as it allows clients to harness the power of LLMs securely.

💡Generative Applications

Generative Applications are software solutions that utilize AI to create new content or insights based on existing data. Kulkarni notes that Elastic's platform enables businesses to develop such applications while maintaining their data's privacy. This capability is crucial in today's digital landscape, where companies are looking to innovate and automate processes without compromising sensitive information.

💡Market Share

Market Share refers to the portion of a market controlled by a particular company or product. Kulkarni emphasizes the importance of capturing market share in the rapidly evolving AI landscape. He discusses how Elastic's strategies, including organic growth and potential acquisitions, are designed to increase their market presence and capitalize on the growing demand for AI solutions.

💡Operating Margins

Operating Margins indicate the percentage of revenue that remains after covering operating expenses. Kulkarni shares that Elastic has maintained growing operating margins, which reflects the company's efficient management of costs relative to revenue growth. This is significant as it demonstrates the company's ability to scale while remaining profitable, a key consideration for investors.

💡Tuck-in Acquisitions

Tuck-in Acquisitions refer to the purchase of smaller companies to enhance the capabilities or market position of a larger company. Kulkarni mentions that Elastic is open to such acquisitions as a strategy to acquire talent and technology that aligns with their mission. This approach allows the company to accelerate its growth trajectory while integrating innovative solutions into its existing offerings.

💡Proprietary Data

Proprietary Data refers to information that is owned by an organization and not generally available to the public. Kulkarni highlights the value of proprietary data as the lifeblood of businesses, stating that it differentiates them in the marketplace. Elastic's approach ensures that clients can use their proprietary data effectively while leveraging AI technologies without exposing that data to external risks.

Highlights

Elastic Data Analytics has experienced a 40% increase in business over the past year.

The company positions itself as a search AI firm, combining the precision of search with AI intelligence.

Elastic works with Fortune 500 companies and government organizations to extract insights from unstructured data.

AI enables businesses to derive value from messy data in unprecedented ways.

Elastic provides a mechanism called Retrieval Augmented Generation to integrate private data into LLMs.

The company ensures that sensitive data remains within the client’s environment for security.

Businesses prefer not to expose proprietary data to external LLMs trained on public information.

Retrieval Augmented Generation helps LLMs provide precise answers without needing access to proprietary information.

Kulkarni highlights the importance of a company's private data in differentiating and competing in the market.

Elastic has invested in machine learning techniques for years, building native models on Elasticsearch.

The company is hiring new talent with machine learning expertise while also considering acquisitions.

Kulkarni mentions the significant cash reserves of about $1 billion for potential acquisitions and organic growth.

Elastic aims to capture market share while maintaining a focus on long-term growth and profitability.

The company has demonstrated consistent revenue growth exceeding expense growth, improving operating margins.

Elastic's strategy includes leveraging both organic and inorganic growth opportunities to expand in the AI space.

Kulkarni expresses optimism about the future potential for growth through market expansion and acquisitions.

Transcripts

tech. Let's talk about

harnessing AI and LMS with the

CEO of Elastic Data Analytics

and Business. That's up about

40% in the past year. Just had

earnings to joining us. Ashutosh

Kulkarni is a CEO at Elastic

Ticker STC. Ash, thanks for

being here on the show with us

this morning. Welcome to the

Schwab Network. Thank you very

much. It's great to be here.

Tell us about your business and

how you guys are dealing with

the avant garde of AI and LMS,

because it seems like you're

kind of right in the middle of

it. Yeah. So, you know, we

describe ourselves as a search

AI company. We bring the

precision of search and the

intelligence of AI and help you

help businesses, companies, big

and small. We work with the

largest fortune 500 companies as

well as government organizations

all around the world to get

insights out of their really

messy, unstructured,

semi-structured data. And as you

know, there's a lot of it. And

AI is providing an amazing

mechanism to really discover and

get value out of that. That

information in ways that were

never possible. And we bring

your private data context to

these large language models. And

through that, we allow you to

bring and build a generative AI

applications for your business.

And that probably is the most

exciting part of our, our,

business today, we just as you

mentioned, announced earnings

for Q4, and our fiscal 24, we

delivered, 20% growth in the

quarter, our cloud business grew

by 32. So a very strong, sort of

beat across all metrics. Your,

growth, too, has been very

consistent, basically at 20,

give or take a couple percent

for about the past year. And it

seems like analysts believe

you're going to be hitting that

going forward. So when you talk

about applying the new

technology into your search

tools and the products that you

provide to big companies, a lot

of big banks, big retailers, is

that coming from in-house

development, how are you putting

that on top of your offerings?

Is it a layer that you're

building, or can you just source

it from what's available

publicly through these language

models? So our core strength is

our search platform and what it

does is it takes all of your

private data. So you know,

businesses, our customers will

bring their private data onto

elastic where it stays within

their environment. So it's

protected. It's safe. You know,

they're able to define all the

privacy controls and so on. And

then through this mechanism

called retrieval augmented

generation, we're able to

provide that internal private

context to these large language

models that could be running,

you know, through the Azure

OpenAI service or through

Google's Gemini models, or

through Amazon Bedrock. You

know, there are so many out

there now, this entire space is

just growing, and you're seeing

models even open up in in open

source. But fundamentally, the

way our customers use us is to

provide that private context

without shipping all your

private proprietary data. And so

they're able to build these

generative applications in a

very privacy first security

forward model. And that's really

the future. You know, that's how

these applications are going to

get built. Is there any risk

that your clients or customers

would say, okay, we used to pay

elastic to do these search

functions for us. Now we can

just plug our data into one of

these models and have them

program it for us. We don't need

to have specialty programmers.

Do you have to stay one step

ahead of that or what do you see

as any potential risk that this

can be done in house? Well,

that's, you know, if you think

about any business out there,

whether you are a retail store,

whether you are a manufacturing

company, a financial services

company, your data is your

lifeblood, your information

about your customers, your

proprietary, you know,

processes. That's what you

differentiate on. That's what

you compete on. Nobody wants to

hand off all of that private

data to these large language

models. And that's the reason

why these large language models,

whether it's from OpenAI or

Google or whoever, they've all

been trained on publicly

available non proprietary

information that's out there on

the internet that you can just

scour and learn from and so on.

They have no concept, they have

no understanding of a

business's, you know, private

proprietary information. And

that's exactly where retrieval

augmented generation and our

search AI platform comes in. We

connect the dots in real time.

So when a question gets asked we

are able to in real time tell

the large language model only

answer this question based on

these few documents. Because of

that, the large language model

does not hallucinate, it knows

exactly how to construct the

answer. It's able to give very

precise answers, and it does not

need to have access to all of

your proprietary information,

which is what all of these

businesses care about. So this

model is the only way businesses

truly expect to be able to do

things going forward. And that's

what's so exciting about the

space that we're in. So do you

have to hire new people that

know how to do this? But then

House or these, programmers that

are are adapting to this because

if we are to believe that this

technology is completely novel

in the last 18 months or year,

how do you build a workforce

around that? Have you needed to

go out and find, kind of new

ways to think about it or is it

able to just kind of adapt and

tweak, with then, the workers

and the expertise you already

have? That's a great question.

And, you know, look, large

language models, foundation

models, as they often refer to,

you know, this is body of work

that's been going on in academia

and in a lot of companies for a

very long time now. ChatGPT

really came onto the stage about

a year and a half ago, and this

was the first real attempt by

any company to get this out,

democratize the space. And since

then, this is, you know, it's

become part of our daily

conversations. But the reality

is that there's been a group of

people who have been working on

these machine learning

techniques for some time now,

you know, within elastic, much

like, you know, with other

companies out there, with a few

other companies out there, we've

been investing in this space for

the last eight, nine years, you

know, we've built ways to make

machine learning models work

natively on data that's stored

in Elasticsearch, which is our

core platform. And we've evolved

that over the years. So we both

have a big pool of organic

talent within the company. We

are growing and we are hiring

people that bring that

additional ML expertise to us.

But we are also a company that

has done, you know, technology

tuck in acquisitions in the

past. So if we find a group of

people that have the same

mission that we are on, that

have that expertise, you know,

we've always looked at, being

acquisitive where it makes sense

, bringing that technology to

help us accelerate our future

forward, but I expect that we'll

do all of those, but feel really

good about our ability to

attract that kind of talent. The

question I was going to ask

maybe just got answered in some

ways, but you had about $1

billion in cash or equivalents

on your balance sheet as of the

last earnings. So I guess if you

do more tuck ins or acquisitions

, there's one use of it. But,

that seems like a pretty chunky

amount of cash when you're,

doing about 300 million a

quarter in revenues. So what do

you want to do with it? Is there

any low hanging fruit, apart

from acquisitions, or is that a

way to put some of that money to

work? Yeah. Fundamentally, what

our investors keep telling us

is, you know, the opportunity is

so large and we believe this in

our hearts. The opportunity is

so massive. We are right in the

sweet spot of AI. And it's

really important that as we look

at the long term, we believe

that we have the opportunity to

build a generational company.

That means we stay focused on

the right way to keep growing,

capture as much market share as

we can. And that means, you

know, deploying this capital

towards not just inorganic, what

we just talked about, but even

our organic growth to make sure

that we can continue to grow

that business and capture more

and more share. We believe that

there's a lot of leverage in our

business, inherent leverage in

our business. So even as we are

growing revenue, what we've

demonstrated in the last many

years is we can grow our revenue

more than we grow our expenses

in a consistent manner. And that

means our operating margins are

non-GAAP. Operating margins have

continued to grow last fiscal

year, we delivered, 11% non-GAAP

operating margin. And the guide

that we've been given for the

current year that we've just

gotten into, is, over 12. So we

believe that, you know, we can

keep doing that in the right way

and keep growing it, in a way

that is going to be right for

long term growth as well as

profitability. Okay ash, thanks

for the explanation, sounds like

we should keep an eye on, some

potential growth via your

market, but also, potential

acquisitions. So appreciate the

insight. Thank you very much.

Absolutely. Ash Kulkarni, CEO at

elastic. All right. Wh