How to Automate Routine Tasks: Python, AI, and Real-World Scenarios

The essential things that will save us from hours of tedious manual work are now available for the first time. They answer the question: how can we finally eliminate all the annoying, repetitive tasks? And they will give us back our valuable time, which we should spend on items that generate revenue for the business. This is the time to start talking about automation — not theoretical abstractions from books, articles and essays, but actual, very practical tools and items that people are using every day to do things like trading through cryptocurrency arbitrage, marketing, office administration, etc. Here are a few actual examples of things people are doing using automation.

«I have been in business for 9 years — I have tried 43 automation tools over that time period. The biggest conclusion is that with the exception of established AI models like ChatGPT, most of the other 'general' tools serve more as a showpiece, examples of false 'benefits' versus legitimate & quantifiable cost savings. Only 'niche-specific, customized' automation solutions provide true, quantifiable cost savings, which is why I have built ASCN.AI — an ecosystem for Web3 — where general models lack the ability to interact with the blockchain in real-time, and that is what our clients require.»

The days of searching for information manually and writing scripts for «every little thing» are no longer relevant. The marketplace is demanding that automation tools be able to process highly specialized data and be able to learn from the data, as well as be able to react to changes immediately.

Introduction to Automating Repetitive Tasks

In other words, automation is the new way to outsource routine and repetitive processes to programs and/or artificial intelligence to free you from performing the same routines over and over. This is not about replacing people, rather freeing them up to work on creative and strategic projects.

Why is automation necessary?

• Because it saves time! The time saved by automating office work can be as high as 30% — no more routine and no more delays.

• With fewer mistakes being made overall (because a program cannot get fatigued or lose its focus), a program produces significantly fewer errors than a human when processing manually.

• Additionally, a single script can handle anywhere from 10 to 1,000 tasks simultaneously without the need for additional staffing or employees.

For example, the Arbitrage Scanner project had its procedure for collecting data from more than 40 exchanges automated. An analyst previously spent six hours per day manually validating spreads, now this task is performed by a Python script in approximately 10 seconds, freeing the human analyst to focus only on decision-making.

Why Automation is Important Right Now

So why is it important to automate now? The speed of the current marketplace renders manual work ineffective. For instance, the opportunity for arbitrage in crypto exists for only a matter of hours, while in marketing, cold leads lose their conversion potential after just five minutes. If you miss your opportunity, you lose your customer.

Automation enables you to achieve the following:

1. Reduce costs. Automated robotic process automation (RPA) provides the potential for operational expenses to be lowered by approximately 40%.

2. Provide immediate reaction times. Automations and automated scenarios have the capability of reacting to any event instantly.

3. Free up employees. Through automation, marketing and technical personnel will no longer be needlessly burdened with repetitive, routine or time-consuming tasks; therefore, report generation will be automated.

Consequently, after the crash of the crypto market by 30% in 2024, those clients that had automated systems or automated services took advantage of the opportunity to profit from their use of spot-futures arbitrage, while others lost their chance to benefit from the lower prices.

The Role of Python in Automation

As an aside, Python is one of the easiest programming languages to learn because of its clear syntax. Both novices and experts can benefit from this.

• Simplicity of use — you are able to write an elementary script in a matter of hours.

• Numerous libraries: software tools that are already available and can be used to create files from various file formats (e.g., text files), work with files/folders via application programming interfaces (APIs), and retrieve data from websites using web scraping techniques.

• Community support: there are many people in the StackOverflow user community who can assist you and provide solutions for errors that you may encounter while using Python programming.

As of this writing, Python is one of the top three programming languages, with minimal direct competition in the automation/scripting realm.

Core Python Libraries for Automation

• gspread and Google Sheets API — working with Google Sheets.

• Selenium — automated testing and web scraping through browser control.

• Pandas — tabular data analysis and filtering/aggregation by different keys.

• OpenPyXL — reading/writing and formatting Excel files.

Examples of Typical Automation Scenarios with Python

Scenario 1: Report from Google Sheets to Telegram

Task: Send a summary of yesterday's sales to the work chat every morning.

import gspread
from telegram import Bot
from datetime import datetime, timedelta

gc = gspread.service_account(filename='credentials.json')
sheet = gc.open('Sales').sheet1
data = sheet.get_all_records()

yesterday = (datetime.now() - timedelta(days=1)).strftime('%d.%m.%Y')
filtered = [row for row in data if row['Date'] == yesterday] 
 
total_orders = len(filtered)
total_revenue = sum(row['Amount'] for row in filtered)
avg_check = total_revenue / total_orders if total_orders > 0 else 0

bot = Bot(token='YOUR_BOT_TOKEN')
message = f"Report for {yesterday}:\nOrders: {total_orders}\nRevenue: {total_revenue} $\nAvg Check: {avg_check:.2f} $"
bot.send_message(chat_id='YOUR_CHAT_ID', text=message)

Scheduled launch via cron (Linux/Mac) or Task Scheduler (Windows).

Scenario 2: Competitor Price Scraping

Task: Check a competitor's prices for 50 products every day.

from selenium import webdriver
from selenium.webdriver.common.by import By
import pandas as pd

driver = webdriver.Chrome()
driver.get('https://competitor.com/products')

prices = []
products = driver.find_elements(By.CLASS_NAME, 'product-item')

for product in products:
    name = product.find_element(By.CLASS_NAME, 'product-name').text
    price = product.find_element(By.CLASS_NAME, 'product-price').text
    prices.append({'Product': name, 'Price': price})

driver.quit()

df = pd.DataFrame(prices)
df.to_excel('competitor_prices.xlsx', index=False)

Such a script saves an analyst up to 3 hours a day.

Scenario 3: Incoming Email Processing

Task: Automatically route emails based on their subject.

import imaplib
import email

mail = imaplib.IMAP4_SSL('imap.gmail.com')
mail.login('your_email@gmail.com', 'your_password')
mail.select('inbox')

status, messages = mail.search(None, 'UNSEEN')
for num in messages[0].split():
    status, msg_data = mail.fetch(num, '(RFC822)')
    msg = email.message_from_bytes(msg_data[0][1])
    subject = msg['subject'].lower()

    if 'refund' in subject:
        # Forward to Finance
        pass
    elif 'technical issue' in subject:
        # Forward to Support
        pass
    elif 'want to buy' in subject:
        # Send Price List
        pass

Using Artificial Intelligence for Automation

An AI system is a computer program that learns from its environment and takes action based on its understanding of contextual relationships within a body of data.

AI can provide a large ancillary benefit to the automation of business processes via:

• Analysis of text (e.g., analysis of reviews to identify positive, negative, or neutral sentiment), identification of major elements contained in textual/nuanced data, etc.

• Classification of data (e.g., separating out customer inquiries based on the type of request).

• Generation of content (e.g., generating product descriptions or responses to customer inquiries).

• Predictive behavior (e.g., forecasting customer preferences or ordering patterns).

ASCN.AI builds artificial intelligence models based on data available in the lecture pool of ASCN.AI, which allows us to deliver highly accurate crypto-related analytics. ChatGPT's crypto models are inherently less accurate than ASCN.AI's because the former does not account for real-time blockchain data.

Popular AI Platforms and Tools

The OpenAI API (GPT-4) allows for generating text through the use of an API while there are also other types of AI technology like LangChain, which is a platform that enables you to build chains of actions powered by AI; and Hugging Face Transformers are for performing text analysis on your local computer. Another example of an AI technology platform is called ASCN.AI NoCode, which is an experimental no-code platform with some AI components that can interface with Telegram, Google Sheets, and CRMs.

An example of ASCN.AI is a Telegram user requesting the current price of a stock and an ASCN.AI agent gathering blockchain data and returning the same data in approximately 10 seconds and/or for example social media analysis, as well as for the top 100 shareholders — returning this data as a result.

How Artificial Intelligence Increases Business Efficiency

• The use of AI to automate customer service: by allowing dialogue between customers and a bot being able to identify the customer's needs vs wants, the time needed to process the customer's needs has decreased from 4 hours to approximately 1 hour.

• The use of AI to personalize marketing: by helping an e-commerce site identify its customers’ purchasing habits, the AI will be able to match those habits with items available on the site, provide them discounts via newsletters, and provide approximately 25% increases in item opens and 15% in purchase conversion.

• Market analysis: the ASCN.AI system contains embedded Ethereum and Solana nodes that scrape the internet for social media information about tokens and exchange information about tokens.

Practical Guide: Managing Routine Tasks

If you want to automate the processes in your business, you should focus on automating those tasks that:

1. happen repetitively;

2. are completed in strict sequence;

3. and have a well-defined and formal procedure.

The use of Python scripts connecting the Facebook and Google Ads API, has enabled a marketing company to save their 5-member team 10 hours a week by automatically exporting their advertising figures to Google Sheets.

Setting Goals and Choosing Tools

Selecting appropriate tools for specific tasks:

• When working with tables: select the Python package Pandas for working with data tables; Google Sheet API, or OpenPyXL for working with Google Sheets.

• When scraping: select one of the following: Selenium, BeautifulSoup, or Scrapy.

• When consuming APIs: select Python’s Requests module.

• When automated no-code: select ASCN.AI NoCode, Zapier, or Make.

• When using AI agents: select any of the following: OpenAI API, LangChain, or ASCN.AI NoCode.

Develop a clear goal. Example: lead posts in the CRM in less than 5 seconds.

Recommendations for Successful Automation

• Select one process/task to automate and test it.

• Documenting the automation process by creating documented procedures.

• Maintain copies of your data for testing.

• Log errors in the automation process.

• Periodically check for updates to APIs.

• Be aware of any limits on number of requests to prevent being blocked.

Examples and Real-World Case Studies

Example of case study: Arbitrage Scanner.

Analyst spends 6+ hours a day collecting prices from 40+ different exchanges for arbitrage analysis. We built a Python program that collects exchange prices and provides a visualization of arbitrage opportunities in real time. The analysis is now done in less than 10 seconds. Read more — detailed analysis of the ASCN.AI case during the Falcon Finance crash

Case study: Flash Crash on October 11.

The Flash Crash on Oct 11 led to markets falling 30% within hours and spot pricing vs futures price divergence of anywhere from 5%-40%. Clients utilizing automation were able to get notifications immediately and get into positions while otherwise, the clients using manual trading processes would have missed this opportunity. Read more — more information regarding the flash crash example

Automation of Office Processes

One manager was able to free up 30 minutes of his time daily by automating exports of CRM reports and sending them through ASCN.AI NoCode to Telegram. Data was made available almost instantaneously across the entire team, enabling faster decision making.

Support was able to automate sorting of 100+ emails by utilizing an AI agent to identify the inquiry category and forward it to the appropriate department. As a result, the average processing time for an inquiry was reduced from 60 minutes to 5 minutes.

Automation in Marketing and Sales

One marketing representative was able to automate the import of leads into Google Sheets and CRM as the result of the website traffic they received. This allowed for a real-time processing of leads and resulted in an 18% increase in conversion from when he would have processed leads manually.

The use of an AI agent to write personalized email newsletters resulted in a jump in open rates from 8% to 23% and a 15% increase in conversion from open to purchase.

Frequently Asked Questions (FAQ)

Conclusion and Recommendations for New Users

Start by automating one task that is recurring. Learn Python and write some simple scripts and develop those. If you aren't a coder then you can use a no-code platform such as ASCN.AI NoCode to build your automation in a visual manner.

Disclaimer

The contents of this article are general in nature and do not constitute investment, legal, or security consultations; therefore when utilizing any AI assistants, you should do so with a prepared mind and knowledge of how specific technology platforms operate.