SECRETS OF

GEN AI SUCCESS
How leaders like Bayer and Siemens Energy use generative AI
to increase revenue, improve productivity and better serve customers
3 The AI reality check
4 Strides toward success

7 Gen AI in action
8 TS Imagine adopts AI at scale to save 30% in costs and 4,000+ hours of effort

10 Siemens Energy’s AI chatbot democratizes access to 700k+ pages of knowledge

11 Bayer’s business teams make better, faster decisions with self-serve insights

13 Sigma increases sales efficiency with prescriptive recommendations

14 Compare Club turns untapped call transcripts into new ways to delight and engage members

15 CARTO simplifies wildfire risk analysis by explaining geospatial data in natural language

16 Terakeet helps customers pinpoint new market opportunities 98% faster

17 Coda’s conversation-based AI platform allows customers to turn data into action

18 A Premier Venture Capital Firm turns third-party data into investment insights

19 Your path to gen AI success

THE AI
REALITY CHECK
When generative AI catapulted to fame in late 2022, hopes were Compound these challenges with the ever-changing AI toolset,
limitless. It was going to supercharge teams by automating drudgery. which seems to grow daily, and AI efforts may start to feel more like
It was going to unlock new business models, unearth revenue futile attempts.
streams and lead to unfettered productivity. It was going to cure But there are still plenty of reasons to be bullish about AI.
diseases and save lives. Gen AI was going to change everything.
Organizations across industries are finding ways to overcome these
Fast-forward to today, and gen AI has changed many things — but hindrances and unleash the potential of AI across their business.
some of its biggest promises are still more aspiration than reality. These leaders are lowering the AI barrier to entry by building on an
While nearly two-thirds of C-suite executives say trust in AI drives already robust data foundation that prioritizes flexibility and helps
revenue, competitiveness and customer success, according to a them stay agile to meet evolving needs. They’re managing costs,
Salesforce survey, Harvard Business Review reports that up to 80% reducing operational burden and ensuring their data stays governed.
of AI projects fail to make it into production.
And they’re achieving these goals and more with Snowflake’s
So why the disconnect? When we examine the obstacles versatile, powerful platform that makes AI easy, efficient
organizations face in bringing their gen AI use cases into production, and trusted.
it’s not hard to understand why many are falling short.

80%
Roadblocks include:
• The cost and complexity of managing multiple AI
of AI projects
platforms — and then moving and duplicating data across
fail to make it
all these infrastructures. into production
• The operational burden of democratizing access to AI,
equipping the right people with access to the right tools
and the right training.
• The governance requirements to ensure that data and
large language models (LLMs) are meeting heightened
security standards.

3
STRIDES TOWARD SUCCESS
Organizations of all sizes and industries are turning to Snowflake’s unified AI and data platform to
make their generative AI goals a reality, whether it’s increasing sales efficiency and accelerating
sales cycles (Sigma) or democratizing access to insights for nontechnical teams as varied as
finance, sales and demand planning (Bayer).
These premier enterprises are succeeding by tapping into the power of Snowflake
Cortex AI, a fully managed service that hosts and serves top-tier LLMs, allowing businesses
to easily and efficiently develop generative AI applications without moving data outside of
Snowflake’s trusted boundary. In doing so, these organizations are building AI applications that
are capturing tangible value — from pinpointing new market opportunities for customers
98% faster (Terakeet), to achieving 30% cost savings through greater efficiency (TS Imagine).

98%
Faster in pinpointing new
market opportunities
30% Cost savings through
greater efficiency

4
THREE KEY USE CASES MAPPING 2. Conversational assistant: Bayer, for example, allows business
THE WAY Thanks to conversational assistants, teams to self-serve insights by
While organizations are exploring a variety enterprises are allowing employees using Cortex Analyst to convert
of new gen AI use cases, there are three to search documents and multimedia natural language to SQL. This
primary areas where customers are making through chat. This capability gives helps teams quickly and easily find
big strides with gen AI: teams intuitive, interactive access valuable information around product
to information and insights while performance, profits and beyond to
1. Text processing at scale:
eliminating the need to dig through files make more strategic decisions.
Organizations are using Snowflake
or multiple applications.
Cortex AI to gain access to industry-
leading, open source LLMs, such
Organizations like Siemens Energy are
as Mistral and Snowflake Arctic,
using Cortex AI to build these powerful
that quickly perform complex tasks
chatbots based on a retrieval augmented
like translation, summarization and
generation (RAG) architecture that
sentiment analysis. This processing
enhances data access and improves
takes place across massive volumes
of text — from summarizing hundreds
employee productivity, while tech
platform Coda uses Cortex AI and
Accelerates
of pages of documents, to calculating
sentiment across thousands of support
its integrated features like Cortex sales cycles
Search (in public preview) to provide
call transcripts, to translating millions of
product reviews — all without leaving
hyperlinked citations that verify the and delivers a better
Snowflake’s secure, governed perimeter.
accuracy of its search results. customer experience
A common way organizations are 3. Automated business intelligence (BI):
generating value with this use case is by Automated BI enables employees to
empowering sales teams. Sigma uses perform ad hoc data analytics without
Cortex AI to turn hundreds of Salesforce writing SQL, which is typically used for
opportunities, support calls and chats generating analytics and interacting
into invaluable reports for sales that with databases. With powerful Cortex
predict a customer’s propensity to AI features like Cortex Analyst (in
public preview) and Snowflake Copilot,
purchase, summarize win/loss insights
and suggest next best actions. And other users leverage intuitive, conversational Enhances strategic
interfaces that streamline data
companies, such as Compare Club, are
also using these capabilities to breathe exploration, visualization and reporting decision-making
newfound value into untapped assets to make data insights available to
like customer reviews, support tickets, everyone — not just technical teams. with self-serve
call transcripts and supplier contracts. data insights
5
THE JOURNEY PAYS OFF • A multibillion-dollar American
These real-world gen AI use cases are equipment rental enterprise uses
delivering real results for companies both big the Mistral LLM in Cortex AI to cost-
and small. In doing so, these organizations effectively analyze thousands of sales
are capturing tangible value from their AI use notes and reduce customer churn.
cases — from Siemens Energy accelerating • A Fortune 50 insurance giant uses
research and development and time to value, Cortex AI to accelerate recruitment
to TS Imagine saving over 4,000 hours of efforts by adding more structure to
manual effort a year while cutting costs by resumes, in addition to building an Eliminates years’
nearly a third. employee chatbot that quickly surfaces
information sprinkled across hundreds
worth of manual
And the enterprise innovation extends far
beyond these pages. Here are just a few other of manuals. effort for
organizations using Cortex AI to achieve their
business goals:
As these businesses illustrate, AI has
enormous potential that can extend into
faster time
• A leading global video hosting platform every team in the enterprise. While the
evolving landscape will continue to spur new
to value
uses LLMs in Snowflake Cortex AI to
generate new insights and sentiment obstacles and opportunities, building on
analysis from video transcripts, Snowflake’s robust AI platform can help your
helping significantly boost conversion organization stay one step ahead to realize
rates by better tailoring communications the promise of AI.
to customers.

Saves over
4,000 hours
a year on tedious
processes

6
GEN AI
IN ACTION
Future-minded leaders across industries are building easy, efficient and trusted gen AI applications
with Snowflake.

See how the organizations in the following pages are using the AI Data Cloud and Cortex AI to achieve their goals —
from better serving customers to gaining competitive advantage.

7
TS IMAGINE ADOPTS AI AT SCALE
TO SAVE 30% IN COSTS
AND 4,000+ HOURS OF EFFORT
By automating email monitoring and streamlining management for support and data tickets,
TS Imagine reduces manual tasks, optimizes data processing and improves efficiency — saving
time and cutting costs.

IMPACT
INDUSTRY Financial technology USE CASE Text processing; conversational assistant • 30% cost savings — without
sacrificing performance — by
using Snowflake vs. other
TS Imagine delivers a SaaS platform for integrated electronic front-office trading, portfolio
industry-leading, pretrained
management and financial risk management. Traders, risk managers and operations executives at
LLM APIs
more than 500 global financial institutions use TS Imagine products so they can better focus on
generating and protecting assets within today’s fast-evolving markets. • Ease of use for data teams to go
from traditional natural language
CHALLENGE processing to generative AI at
Before Snowflake, TS Imagine’s homegrown data solutions required redundant, manual work. Every year, members scale in only 6 months
of the data management team spent thousands of hours reading more than 100,000 emails to monitor for
notifications from their data providers. These critical emails detailed upcoming product changes, any of which
• Massive time savings for
could ultimately impact TS Imagine’s clients. Missing even a single email could lead to a downstream product employees, from over 4,000 hours
outage — and very unhappy customers. spent on manual email monitoring
Another pain point was prioritizing the more than 60,000 support and data tickets that the customer team tasks to near-instant, AI-powered
received annually. Since TS Imagine comprises two legacy firms founded in the 1980s, finding internal information categorization for just one
about the combined company post-merger was also inefficient and time-consuming for current employees and use case
new hires, with knowledge spread out in far too many places. Finally, employees had to peruse dense terms and
conditions documents published by derivatives exchanges about thousands of financial instruments to produce
structured data for TS Imagine’s trading and risk management products. Read the full story

8
SOLUTION
With Snowflake, I can empower
TS Imagine has operationalized AI at scale with Snowflake Cortex AI. Instead of employees
spending over 4,000 hours a year on an error-prone sorting process, TS Imagine has smart people to bring AI to life in
automated email intake, using Cortex AI to delete duplicate or nonrelevant messages, and
create, assign, prioritize and schedule JIRA tickets for each email. The team hasn’t missed
one place, even if they’re not tech
a notification since implementing the new process in December 2023. And those 60,000+ experts. Cortex AI is a one-stop
customer support and data tickets? Cortex AI classifies them by sentiment, complexity,
urgency, clarity, uniqueness and impact so client managers and support team members can shop. It scales, it’s easy, and the
resolve customer-facing issues much faster. data stays 100% in Snowflake’s
TS Imagine uses Cortex AI to improve internal efficiencies, too. In-depth knowledge about
TS Imagine’s products — found in tens of thousands of articles, documents and support
secure environment.”
tickets within the company’s knowledge base — is now all available via LLMs in Cortex AI, —THOMAS BODENSKI
including Snowflake Arctic, Llama and Mistral. Employees find info about the company fast
COO & Chief Data & Analytics Officer,
by simply asking questions of a chatbot, powered by Streamlit in Snowflake. And rather than
painstakingly reading terms and conditions within fact sheets about financial instruments, TS Imagine
teams rely on Snowflake Arctic to transform such PDFs into easily digestible structured data,
which then automatically feeds into TS Imagine’s products.

4,000+ HOURS SAVED

30% COST SAVINGS

9
SIEMENS ENERGY’S AI CHATBOT
DEMOCRATIZES ACCESS TO 700K+ IMPACT
PAGES OF KNOWLEDGE • Information from 700K+ pages
of dense documents summarized
With Snowflake Cortex AI and Streamlit in Snowflake, Siemens Energy built an AI chatbot to quickly
and available to teams, which
surface and summarize over half a million pages of internal documents, helping the R&D team maximize
productivity and effectiveness while accelerating decision-making and competitiveness.
accelerates R&D, innovation and
time to market
• ~25 highly specialized R&D
engineers now have near-instant
INDUSTRY Manufacturing USE CASE Conversational assistant access to data and insights
• Hours of time saved — amounting
A global leader in the energy business, Siemens Energy’s innovative solutions contribute to to years’ worth of effort — by
around one-sixth of the world’s energy production. Half of Siemens Energy’s portfolio is based eliminating the need to manually
on technology pertinent to a decarbonized energy world, demonstrating the company’s deep search for data
commitment to sustainable energy technology.

CHALLENGE Watch their presentation

Siemens Energy has a deep knowledge base contained in more than 700,000 pages of proprietary research and
development documents. The company wanted to make the valuable information in these paper documents
available to its globally distributed research and development (R&D) team — but reading all this information would
take someone a staggering four years. The challenge was to not only digitize the information, but to then make This chatbot, built with Snowflake Cortex
it more accessible to teams so they could use the wealth of knowledge to accelerate discovery, collaborate more and in combination with our comprehensive
effectively, boost decision-making and fuel innovation. citizen development approach, has been a
game changer for Siemens Energy, particularly
SOLUTION in the realm of fast and economic data
democratization. It has optimized our resource
Applying a comprehensive citizen development approach, Siemens Energy built an intuitive document Q&A allocation, saving many manual hours spent
chatbot in Snowflake that democratized knowledge discovery and collaboration in a secure, highly economical looking for information. The ability to unlock
way. Developed in a very short period of time, this state-of-the-art chatbot uses a retrieval augmented generation and democratize the data hidden in our
(RAG)-based architecture — built using Streamlit in Snowflake and Snowflake Cortex AI — that made the internal data treasure trove has given us a distinct
company documents securely available to off-the-shelf large language models without the need for training, fine competitive edge. It’s intuitive, it’s innovative,
tuning or moving data outside Snowflake’s governed platform. and it’s driving us forward in our mission —
Whereas they had to previously spend hours sifting through and reading research papers, researchers now including achieving high scalability for further
simply use the chatbot to quickly identify relevant research excerpts with accompanying hyperlinks to the source use cases already in progress.”
material. This breakthrough helps Siemens Energy save years’ worth of manual searching while ultimately speeding —TIM KESSLER
time to market and sharpening the company’s competitive edge. Head of Data, Models & Analytics, Siemens Energy

10
BAYER’S BUSINESS TEAMS
MAKE BETTER, FASTER DECISIONS
WITH SELF-SERVE INSIGHTS
Bayer’s Consumer Health North America team converts natural language queries into accurate and
actionable SQL, helping nontechnical users — from customer finance to sales to demand planning
— quickly analyze data for faster decision-making, without relying on data teams.

IMPACT
• More strategic decision-
making by giving business
teams the ability to ask
Healthcare and life sciences Automated BI
INDUSTRY USE CASE
questions of Bayer’s curated
data sets in natural language
Guided by the mission “Health for all, Hunger for none,” Bayer is a global life sciences enterprise • Insights at scale for users
that designs products and services to help tackle some of the world’s biggest challenges while across the enterprise — from
serving the most essential human needs of health and nutrition. sales to customer finance to
CHALLENGE demand planning
Businesses like Bayer move fast, especially when partnering with third-party retailers and suppliers as part of • Better productivity for both
their go-to-market strategy. To make quick, informed decisions about the business, teams across the company technical and nontechnical
— from customer finance to sales to demand planning — need near real-time, accurate insights from their data. departments by reducing
Yet most teams only had access to this data via dashboards, which often resulted in information overload for reliance on data teams
users and cumbersome, time-consuming development for data teams. With limited coverage and flexibility, these and enabling self-serve
dashboards forced users to manually dig through and interpret data, which hindered teams from quickly extracting
data access
insights, increased analytics inconsistencies and introduced the potential for human error.

Watch their presentation

11
SOLUTION
The core of this capability is high-
Leveraging Cortex Analyst to abstract database interactions by converting natural language to
SQL, Bayer now empowers its business users to easily self-serve insights through an intuitive quality responses to a natural
Streamlit in Snowflake chatbot interface. Instead of having to comb through dashboards or
request support from data and analytics teams, departments like customer finance, sales and
language query on structured data,
demand planning use natural language to seamlessly interact with analytical data stored in used in an operationally sustainable
Snowflake tables. This helps them find answers faster while unearthing richer, actionable
insights in easy-to-understand terms and visuals. For example, a customer finance user can way. This is exactly what Snowlake
request a tabular report of customer profit and loss statements by brand and month, or ask Cortex Analyst enables for us.
which products have the best and worst performance by net sales at a given retailer. Using
Cortex Analyst’s text2SQL capabilities, the chatbot returns accurate, reliable results in near What I’m most excited about is
real time to help teams better understand the business and make strategic decisions — no
technical expertise or IT support required.
we’re just getting started, and we’re
looking forward to unlocking more
value with Snowflake Cortex AI.”
—MUKESH DUBEY
Product Management and Architecture Lead,
Bayer U.S.

12
SIGMA INCREASES SALES
EFFICIENCY WITH PRESCRIPTIVE
RECOMMENDATIONS
Armed with analyses like an account’s propensity to purchase and “why we win” summaries, IMPACT
go-to-market teams save time, accelerate sales cycles and deliver a more tailored experience to • Reduced time and effort for
prospects and customers. sales teams to identify the
most promising prospects by
using AI to predict an account’s
propensity to purchase
INDUSTRY Technology USE CASE Text processing and summarization
• Better understanding of sales
motions and successes by
Sigma’s powerful business intelligence platform scales billions of records using spreadsheets, equipping teams with win/loss
SQL, Python and AI — without compromising speed and security. Its solution helps users analyze summaries
information via dashboards, then gain access to the underlying data in an intuitive, spreadsheet-
• Access to rich, easy-to-digest
like interface. This interface acts as a frontend to cloud data platforms like Snowflake and provides
insights around customer
the average spreadsheet user with a simple way to interact with LLMs in Cortex AI.
preferences, needs and
CHALLENGE sentiment by summarizing
Sigma helps teams of all sizes, from startups to Fortune 500s, better understand and interact with their data. hundreds of Salesforce
But to deliver on this goal — and continue helping more companies — Sigma’s own internal teams require a deep opportunities, account team
understanding of both their customers and prospects. Sales teams needed a scalable way to easily access data notes and support calls and
while pinpointing the most promising leads so they could prioritize efforts and maximize their time. At the same transcripts
time, Sigma’s go-to-market and operations teams needed a more reliable way to stitch together account details
rather than relying on individuals’ notes, which could be incomplete, inconsistent or prone to human error.
Watch their presentation
SOLUTION
Sigma now uses Cortex AI to fuel productivity and insights for teams across the business. Through Cortex
AI, Sigma taps into the power of LLMs like Mistral and Snowflake Arctic to predict a prospect’s propensity to
purchase and equip sales teams with suggested next-best actions to accelerate the sales cycle. Business teams Snowflake’s Cortex AI has immediately
across Sigma can also ask questions of the company’s data to better understand customer needs on a higher level, unlocked new workflows for our team. When
such as “why we win” summaries and customer sentiment analysis from hundreds of Salesforce opportunities and combined with Sigma’s Input Tables in our data
support calls and chats. Available within an internal Sigma app, these insights help sales and operations teams applications, we’re turning these model outputs
focus their efforts, provide better service and understand customer needs on a deeper level — benefits that have into real value faster than ever before.”
made this innovation wildly popular among Sigma’s teams. —JAKE HANNAN,
Head of Data, Sigma

13
COMPARE CLUB TURNS UNTAPPED
CALL TRANSCRIPTS INTO NEW WAYS
TO DELIGHT AND ENGAGE MEMBERS
Compare Club uses Cortex AI to efficiently summarize hundreds of thousands of pages of call
transcripts, extracting valuable insights that improves member experiences and team productivity
while keeping customer data secure.

IMPACT
INDUSTRY Financial services USE CASE Text processing
• Enhanced member experiences
thanks to fresh insights from
Compare Club helps millions of Australian consumers make more informed purchasing decisions on thousands of lengthy call
products and services across health and life insurance, energy, home loans and more. transcripts
By treating data as a valuable tool, Compare Club experts make it easier and faster for customers
• Greater productivity and
to sign up for the most appropriate product and get more value from their budgets.
effectiveness for teams like
CHALLENGE sales and support with easily
Providing an exceptional, personalized experience to customers is critical for Compare Club — especially for
accessible, comprehensive
returning members, who are more likely to make a purchase. Customer calls are an important vehicle to deliver customer details
this experience, yet complex details from these conversations were not always recorded in the company’s • Peace of mind for sensitive
CRM, making it difficult to use this information in future calls. This oversight caused frustration and prevented
information by keeping customer
Compare Club’s experts from accessing the complete customer history they needed to provide the most effective
data within Snowflake’s secure,
recommendations. While call transcripts contained a trove of valuable insights, the team would have had to
manually sift through thousands of transcripts — each exceeding 100 pages — which wasn’t feasible from a time governed perimeter
and bandwidth perspective.
Compare Club wanted to use generative AI to summarize these transcripts, but document length presented
Learn how they built it
challenges to various LLMs’ context window limits. And running a summarization task for every call meant having
to move large volumes of sensitive call recordings into a third-party service, which would require validation and
additional security approvals.

SOLUTION
Compare Club quickly overcame these challenges by using Cortex AI to run LLMs securely inside Snowflake,
eliminating the need to move data while easily running both preprocessing and LLM tasks with a bit of SQL and
Python. Now, Compare Club efficiently equips business teams with valuable insights extracted from hundreds
of thousands of transcript pages, including details like customer goals, needs, objections, loyalty, history and
enthusiasm. These nuances help Compare Club teams — from sales to support to customer success — better
serve and engage repeat members to improve their experience and retention.

14
CARTO SIMPLIFIES WILDFIRE
RISK ANALYSIS BY EXPLAINING
GEOSPATIAL DATA IN NATURAL
LANGUAGE
By turning complex geospatial data into clear, natural language explanations and graphics, CARTO
makes it easier for nontechnical users to understand complex predictions like wildfire risk scores.
IMPACT
• More transparency by translating
complex geospatial data into
INDUSTRY Technology USE CASE Text processing and summarization
natural language and interactive
Streamlit in Snowflake
As the leading location intelligence platform, CARTO aims to make geospatial analytics accessible dashboards
to all audiences. By presenting data visually, the company helps organizations in industries like • Better understanding for
urban planning, environmental management and business analytics drive better outcomes — from nontechnical users seeking
more efficient delivery routes to better behavioral marketing. to understand complex data
CHALLENGE questions like how insurance
premiums are calculated
CARTO helps users create geospatial map dashboards, models and workflows, which are essential for visualizing
complex spatial data and uncovering geographic patterns that inform strategic decision-making. But when a data • Greater productivity for data
scientist creates and maps predictions from a particular machine learning model, nontechnical users often find scientists by automating
it challenging to understand the rationale. For example, if a homeowner receives a letter informing them that explanations that accompany
insurance premiums are increasing due to a fire risk index, they want to understand the cause. This also means
their analysis
that insurance companies are under mounting pressure to provide transparency and clear explanations around
how they’re rating different locations.

Learn how they built it

SOLUTION
With Cortex AI, CARTO translates complex metrics used to decide which ML model to use into a more accessible,
easier-to-understand explanation for a nontechnical user. Building on the wildfire example, CARTO created a
composite index using geospatial data to rank the wildfire risk of a given area on a scale of “no risk” to “very high.”
Using Cortex AI functions, CARTO then summarizes and explains this map data in natural language for users, such
as showing that an average annual temperature increase of three degrees has elevated fire risk in their region. To
make this data even more accessible, CARTO visualizes the results in an intuitive Streamlit in Snowflake dashboard
that allows users to interact with the data.

15
TERAKEET HELPS CUSTOMERS
PINPOINT NEW MARKET
OPPORTUNITIES 98% FASTER
Using advanced LLMs to assess market opportunities faster, Terakeet has slashed analytics
time while enhancing security by eliminating the need to move sensitive data.

IMPACT
• 98% faster TAM calculations —
INDUSTRY Technology USE CASE Text processing
from 48 hours to 45 minutes
• Increased customer satisfaction
Terakeet is the preferred Owned Asset Optimization (OAO) partner for Fortune 500 brands by identifying untapped market
seeking to connect with customers and grow their business online. Terakeet helps brands optimize opportunities more efficiently
and unify their digital assets, such as company values, websites and market positioning, to meet
consumers as they search for solutions.
• Enhanced security by eliminating
the need to move data and
CHALLENGE building gen AI on Snowflake’s
Terakeet helps its customers make total addressable market (TAM) estimations to understand potential revenue governed platform
opportunities for their products or services, and effectively target untapped markets or customer segments.
These assessments help brands understand the total addressable market of keywords and topics most relevant
to their target audience, so they can tailor their marketing and development strategies accordingly. To build these
Snowflake Cortex AI has changed how we
assessments, Terakeet starts from a client’s online properties. After pulling the client’s and competitors’ search
extract insights from our data at scale using the
landscape, the full set of results contains keywords and topics that may not be relevant. Prior to the TAM project,
power of advanced LLMs. Our teams can now
refining these results required a variety of methods, including tokenization,term matching and topic modeling that
quickly and securely analyze massive data sets,
used an older natural language processing model. But these methods were cumbersome and time consuming,
unlocking strategic insights to better serve our
requiring significant manual effort.
clients. We’ve reduced our processing times by
64x with the power of Snowflake’s new
SOLUTION AI features.”
The text processing capabilities of Snowflake Cortex AI let Terakeet modernize and accelerate the TAM assessment —JENNIFER BRUSSOW
process. Cortex AI has prebuilt task-specific functions that use LLMs behind the scenes to perform complex tasks Director of Data Science, Terakeet
like translation and summarization. With easy access to embedding functions in Cortex AI, the data science team
transforms textual data into numerical vectors — or embeddings — which can identify data patterns and trends
like semantic similarity. The team now performs these calculations directly in the Snowflake platform, which has
eliminated the need to move data to third-party services outside of Snowflake’s secure perimeter. This has not only
resulted in significant time savings, reducing the assessment process from 48 hours to 45 minutes, but it has also
enhanced data security. With such marked results, Terakeet’s marketing and sales operations teams now use Cortex
AI’s capabilities for all of their clients and potential customers.

16
CODA’S CONVERSATION-BASED AI
PLATFORM ALLOWS CUSTOMERS TO
TURN DATA INTO ACTION
Coda Brain helps customers interact with their data through a conversational assistant,
providing easy, actionable access to both structured and unstructured data.
IMPACT
• Increased enterprise efficiency
and productivity
Technology Conversational assistant
INDUSTRY USE CASE
• Verifiability of knowledge with
citations through Cortex Search
Coda’s technology transforms how businesses interact with and leverage their data. Coda’s • Enhanced security and
unified platform blends the flexibility of documents, the structure of spreadsheets, the power of governance with Snowflake’s
applications and the intelligence of AI into a single interface. granular data permissions
CHALLENGE and RBAC
Coda’s unified platform serves as a central hub of knowledge for customers. To empower companies with the
ability to extract actionable insights from their vast amounts of data, Coda wanted to build conversational services Watch their presentation
on top of its platform. However, traditional enterprise tools are not equipped to handle the unique challenges
posed by enterprise data. The services needed to navigate complex permission structures, deliver precise and
verifiable answers, and adhere to stringent enterprise-level security and governance requirements.
Snowflake Cortex AI forms all the core
SOLUTION building blocks of constructing a scalable,
Built on Snowflake and using Snowflake Cortex AI, Coda Brain is a turnkey RAG platform that works as a secure AI system. Coda Brain uses almost
conversational assistant for enterprise customers. Coda Brain helps users ask questions of their enterprise data every component in this stack: The Cortex
— structured or unstructured — and get reliable results to turn data into action across the enterprise. Teams no Search engine that can vectorize and index
longer need to spend countless hours searching for enterprise information, and Coda Brain uses Snowflake’s unstructured and structured data. Cortex
robust role-based access control (RBAC) to ensure data stays secure. To minimize hallucinations and increase the Analyst, which can magically turn natural
accuracy of search results over text-based documentation, Coda uses Cortex Search and provides citations to the language queries into SQL. The Cortex LLMs
data contained in responses. And by using Cortex Analyst to turn natural language text into analytical answers via that do everything from interpreting queries to
SQL and accessing structured data, Coda further democratizes access to insights. reformatting responses into human-readable
responses. And, of course, the underlying
Snowflake data platform, which can scale and
securely handle the huge volumes of data being
pulled into Coda Brain.”
—SHISHIR MEHROTRA
CEO, Coda

17
A PREMIER VENTURE CAPITAL
FIRM TURNS THIRD-PARTY DATA
INTO INVESTMENT INSIGHTS
This financial firm uses LLMs to securely analyze and summarize large volumes of data,
providing investment insights without the need for additional infrastructure or resources.

INDUSTRY Financial services USE CASE Text processing and generation IMPACT
• Cutting-edge LLMs available —
Focusing on software and internet investments, this premier American venture capital firm has zero new vendor
been an early investor in several leading technology companies. onboarding required

CHALLENGE • No dependency on engineers to

access, orchestrate and deploy
This leading firm wanted to explore leveraging AI and LLMs to improve efficiency and productivity for several
functions in the business, including marketing, research, portfolio impact and human resources. But the firm’s
pipelines using LLMs
small technology team was made up of just two data scientists and two full-stack engineers. The company lacked • Instantly available, security-
the resources to spend time vetting new vendors or to access, orchestrate and deploy pipelines using LLMs. Given vetted environments to process
the confidential nature of the business, the team also had to ensure the highest levels of security and governance
proprietary marketing and
for its data and AI.
sales database
SOLUTION
Using its existing, production-ready stack of data build tool (dbt) and Snowflake, the data sciences team turned
Snowflake Cortex AI has made our database
to Snowflake Cortex AI to create value for the business without onboarding a new vendor or spending money
more robust with AI-enabled categorizations
on additional infrastructure, resources or upskilling. With access to industry-leading, open source LLMs within
and classifications. We’ve improved our
Snowflake Cortex AI, the team now runs text analytics across millions of records to gain deeper insights into both
due diligence, spun up competitive market
public and private companies. The firm has streamlined sourcing for investment teams by better categorizing and
dashboards more quickly, and enhanced
summarizing company data, such as company type and job title classifications, as well as providing investment
similar company and talent searches. Our data
teams with LLM-generated recommendations for similar companies by measuring statistics like web traffic growth.
scientists now use Cortex AI to get some pretty
exciting results with minimal effort and time
investment.”
—HEAD OF DATA PRODUCTS
Venture Capital Firm

18
YOUR PATH TO
GEN AI SUCCESS
The customer stories in this book are just the
beginning of what gen AI can do for organizations using
Snowflake’s easy, efficient and trusted AI data platform.
Whether you’re just starting to experiment with
generative AI or already have your first AI applications
in production, Snowflake enables you to extend
the benefits of AI to your entire organization by
streamlining development, reducing complexity and
helping secure your data every step of the way.
Discover how Snowflake can help your organization be
more innovative and competitive through generative AI.
Visit Snowflake’s Gen AI webpage for more information
and hands-on resources to help you get started.

19
 ABOUT SNOWFLAKE
Snowflake makes enterprise AI easy, efficient and trusted. Thousands of companies around the globe, including hundreds of the world’s largest,
use Snowflake’s AI Data Cloud to share data, build applications, and power their business with AI. The era of enterprise AI is here.

Learn more at snowflake.com (NYSE: SNOW)

© 2024 Snowflake Inc. All rights reserved. Snowflake, the Snowflake logo, and all other Snowflake product, feature and service names mentioned herein
are registered trademarks or trademarks of Snowflake Inc. in the United States and other countries. All other brand names or logos mentioned or used
herein are for identification purposes only and may be the trademarks of their respective holder(s). Snowflake may not be associated with, or be sponsored
or endorsed by, any such holder(s).