Use cases

Use Upgini for Marketing and CRM

New customers

Look-alike

Customer
segmentation

Fake leads

LLMs to increase
the Conversion Rate
of CRM campaigns

LLMs to decrease
the Churn Rate
Find new customers

Find new customers

75% match rate for Phone numbers
40% match rate for Emails
99% match rate for IP-addresses
Use case description
Find new customers from 200+ external data sources that connected to Upgini. Contact them and make a direct proposal to involve them in your business.
Why does it matter?
Finding new customers is crucial for every business for several reasons:
  1. Revenue Growth: New customers contribute to revenue growth. They bring in new sales, which directly impacts the bottom line of the business.
  2. Business Expansion: New customers can help a business expand into new markets.
  3. Sustainability: Finding new customers ensures the sustainability of your business.
  4. Brand Awareness: New customers can increase the visibility and awareness of your brand.
How to find new customers with Upgini
  1. Obtain a training dataset with customer ID and label. Upgini supports the following customer IDs: phone number, IP address, and email. The label is a mark of your customer or not. You will need some examples of IDs that are not your customers. If you don't know how to get them, contact Upgini, and we will help you.
  2. Search for new relevant features for your task: launch FIT method of Upgini and get a list of relevant features from connected data sources.
  3. Make a test enrichment: launch TRANSFORM method of Upgini and enrich your dataset with only new relevant features. This dataset could consists any phone number or any IP-address.
  4. Use Upgini in production: Upgini will transfer relevant features for the all Phone numbers, Emails and IP-addresses to your DWH. You may use this data directly in your DWH.
Try Upgini to find new customers
Just copy the code and send it to your analyst to repeat!

Upgini will quickly generate a quality report. Ask your analyst to send it to you to check the result.

Copied
%pip install -Uq upgini

# read labeled data
df = pd.read_csv("train_data_set.csv")

from upgini import FeaturesEnricher, SearchKey

# map data
enricher = FeaturesEnricher(
    search_keys={'Date': SearchKey.DATE,
                'Phone_number': SearchKey.PHONE,
                'IPv4': SearchKey.IP})

# launch fit step
enricher.fit(df[['Date','Phone_number','IPv4']], 
             df['client_label'])

# launch enrichment step
df_enriched = enricher.transform(df)
Look-alike

Look-alike

75% match rate for Phone numbers
40% match rate for Emails
99% match rate for IP-addresses
Use case description
Look-alike is used to identify new users that are similar to a given set of user. It is commonly used in applications such as recommendation systems, fraud detection, and customer segmentation. The goal is to find patterns or similarities in the data that can be used to make predictions or gain insights. Look-alike is also often used in CRM and direct marketing to find an audience that is similar to the target one.
Why does it matter?
  1. A look-alike may increase the effect from your CRM and Direct Marketing campaigns two to three times.
  2. Look-alike may be used for marketing research. Don't spend money to get insights from interviews, use external data directly.
How to solve a Look-a-like task with Upgini
  1. Obtain a train dataset with customer ID and label. Upgini supports the following customer IDs: phone number, IP address, and email. Label is a segment ID or any target variable.
  2. Search for new relevant features for your task: launch FIT method of Upgini and get a list of relevant features from connected data sources.
  3. Make a test enrichment: launch TRANSFORM method of Upgini and enrich your dataset with only new relevant features. This dataset could consists any phone number or any IP-address.
  4. Use Upgini in production: Upgini moves relevant features for the all Phone numbers, Emails and IP-addresses to your DWH. You may use this data directly in your DWH.
Customer segmentation

Customer segmentation

75% match rate for Phone numbers
40% match rate for Emails
99% match rate for IP-addresses
Use case description
Customer segmentation is a strategy used by businesses to divide their customer base into distinct groups or segments based on certain characteristics or behaviors. This allows businesses to better understand their customers and tailor their marketing efforts, products, and services to meet the specific needs and preferences of each segment.
Why does it matter?
By effectively utilizing customer segmentation, businesses can improve customer satisfaction, increase sales, and drive overall business growth.
How to solve a Look-a-like task with Upgini
  1. Obtain a train dataset with customer ID and segment ID. Upgini supports the following customer IDs: phone number, IP address, and email.
  2. Search for new relevant features for your task: launch FIT method of Upgini and get a list of relevant features from connected data sources.
  3. Make a test enrichment: launch TRANSFORM method of Upgini and enrich your dataset with only new relevant features. This dataset could consists any phone number or any IP-address.
  4. Use Upgini in production: Upgini moves relevant features for the all Phone numbers, Emails and IP-addresses to your DWH. You may use this data directly in your DWH.
Prevent fake leads

Prevent fake leads

75% match rate for Phone numbers
40% match rate for Emails
99% match rate for IP-addresses
Use case description
Ad fraud is any attempt to defraud advertisers for financial gain. Scammers often use bots to carry out ad fraud. Scammers could generate fake leads for you that could never be contacted.
How Upgini could help to detect fake leads?
Upgini can help you arbitrage leads from external systems. It can score leads and detect which lead is a fake, so you don't waste time and money contacting it.
How to use Upgini for fake leads classification?
  1. Obtain a training dataset for determining fake leads that includes customer ID and label. Upgini supports the following customer IDs: phone number, IP address, and email. The label indicates whether a lead if fake or not.
  2. Search for new relevant features for your training dataset: launch the FIT method of Upgini and receive a list of relevant features from connected data sources.
  3. Make a test enrichment: launch the TRANSFORM method of Upgini and enrich your dataset with only the new relevant features.
  4. Use Upgini in production: Upgini transfers relevant features for all phone numbers, emails, and IP addresses to your DWH. You can use this data directly in your DWH.
Fraud Score
If you don't have your own Fake leads detection system or you don't have a training dataset, you may use the ready-made Upgini Fraud Score. The Upgini Fraud Score is an informational tool that helps you gauge the risk involved with leads from ads before processing them. This is done by identifying traits and historical trends associated with suspicious behavior and fraudulent contacts that you receive from ads.
Use GPT in Upgini to increase the Conversion Rate of your CRM campaigns

Use GPT in Upgini to increase the Conversion Rate of your CRM campaigns

+20% in accuracy to openai library
+15% in conversion rate of your CRM campaigns
Use case description
Companies engaging with their current or prospective customers through phone or chat frequently encounter issues with low conversion rates. These issues stem from the unique characteristics of the call center utilized for client interactions. Not all call center employees may possess the necessary qualifications to accurately document the outcome of each client communication or remember to schedule callback reminders at a time suitable for the client. Furthermore, some operators may commit errors during the conversation.
How to solve this problem with Upgini?
Upgini has a function to extract knowledge from conversation texts of your call center and a customer, leveraging the extensive knowledge base of Large Language Models. Try to utilise additional knowledge from conversation texts:
  1. To identify customers who, despite negative feedback from call center operators, are interested in your product and require a callback.
  2. To recognize customers who, due to high levels of dissatisfaction, no longer need to be contacted.
  3. To evaluate the quality of work performed by call center operators.
How to use Upgini for this task?
  1. Obtain a train dataset with customer ID, conversation text, and target label. The customer IDs that Upgini supports for this task are phone numbers and emails. Use manually labeled events such as "client asked for callback", "client asked to never be called again", or "operator made a mistake in the speech" as target labels.
  2. Launch GPT feature generation in Upgini for your task: launch FIT method with parameter generate_features and you'll get a list of relevant features from connected data sources and relevant features extracted from conversation texts.
  3. Make a test enrichment: launch TRANSFORM method of Upgini and enrich your dataset with only new relevant features.
  4. Use Upgini in production: Upgini provides relevant GPT features through online API
Try to test Upgini to increase Conversion Rate with GPT
Just copy the code and send it to your analyst to repeat!

Upgini will quickly generate a quality report. Ask your analyst to send it to you to check the result.

Copied
%pip install -Uq upgini

# read labeled data
df = pd.read_csv("gpt_train_dataset.csv")

from upgini import FeaturesEnricher, SearchKey

# map data
enricher = FeaturesEnricher(
    search_keys={'Date': SearchKey.DATE,
                'phone_number': SearchKey.IP},
	generate_features=['text1', 'text2']
)

# launch fit step
enricher.fit(df[['Date','phone_number', 'text1', 'text2']], 
             df['call_back_required'])

# launch enrichment step
df_enriched = enricher.transform(df)
Use GPT in Upgini to decrease the Churn Rate

Use GPT in Upgini to decrease the Churn Rate

+20% in accuracy to openai library
+15% in conversion rate of your CRM campaigns
Use case description
Companies that interact with their current customers via phone or chat for customer retention often face issues with low retention. These challenges stem from the unique characteristics of the call center used for customer interaction. Not all call center employees may be qualified to accurately document the outcome of every interaction or remember to schedule call reminders at a convenient time for the customer. Additionally, some operators may make mistakes during calls.
How to solve this problem with Upgini?
Upgini has a function that extracts knowledge from conversation texts between your call center and a customer, leveraging the extensive knowledge base of Large Language Models. Attempt to utilize additional knowledge from conversation texts.
  1. This can be used to identify customers that indeed have a high likelihood of churn. Avoid wasting time and promotions on clients who aren't genuinely on the brink of churn.
  2. Also, it can identify customers that require a callback, even if the operator has categorized them under "do not call back".
  3. Lastly, it can be used to assess the quality of work carried out by the call center operators.
How to use Upgini for this task?
  1. Obtain a train dataset with customer ID, conversation text, and target label. The customer IDs that Upgini supports for this task are phone numbers or emails. Use manually labeled events such as "client have churned"as target labels.
  2. Launch GPT feature generation in Upgini for your task: launch FIT method with parameter generate_features and you'll get a list of relevant features from connected data sources and relevant features extracted from conversation texts.
  3. Make a test enrichment: launch TRANSFORM method of Upgini and enrich your dataset with only new relevant features.
  4. Use Upgini in production: Upgini provides relevant GPT features through online API
Try to test Upgini to decrease Churn Rate with GPT
Just copy the code and send it to your analyst to repeat!

Upgini will quickly generate a quality report. Ask your analyst to send it to you to check the result.

Copied
%pip install -Uq upgini

# read labeled data
df = pd.read_csv("gpt_train_dataset.csv")

from upgini import FeaturesEnricher, SearchKey

# map data
enricher = FeaturesEnricher(
    search_keys={'Date': SearchKey.DATE,
                'phone_number': SearchKey.IP},
	generate_features=['text1', 'text2']
)

# launch fit step
enricher.fit(df[['Date','phone_number', 'text1', 'text2']], 
             df['churn'])

# launch enrichment step
df_enriched = enricher.transform(df)

Make ML&AI models happier
with more relevant data

No credit card required. No time limit on Free plan

Get started
Use cases
Fraud preventionMarketing and CRMDemand forecastProduct managementPersonalizationRisk managementLead scoring
Resources
DocsTutorials & Publications
PricingSell your data
for AI & ML applications
aicpa soc
SOC 2
© 2022-2024 Upgini  ·  Designed in UAE, build worldwide