Demo AI Agents

Explore how demo AI agents can streamline your workflows, automate repetitive tasks, and provide intelligent assistance across your business. Our configurable agents can handle customer support, data research, content drafting, and internal process automation, all while adapting to your specific rules and tone of voice. Use them to prototype new ideas quickly, validate use cases, and demonstrate AI capabilities to your team or clients. With clear reporting and easy controls, you stay in charge while the agents do the heavy lifting in the background.

AI Resume Shortlisting Agent (Keyword + AI Scoring) 

import pandas as pd

from sklearn.feature_extraction.text import TfidfVectorizer

from sklearn.metrics.pairwise import cosine_similarity

# -----------------------------

# STEP 1: Sample Resumes (TEXT)

# -----------------------------

resumes = {

"Resume_1": """

Data Analyst with 3 years experience.

Skilled in Python, SQL, Pandas, and data visualization.

Experience in Machine Learning and statistics.

""",

"Resume_2": """

Software Developer with Java and web development skills.

Experience in HTML, CSS, JavaScript.

Basic knowledge of databases.

""",

"Resume_3": """

Data Scientist with strong Machine Learning background.

Skilled in Python, deep learning, NLP, and data analysis.

Experience with TensorFlow and visualization tools.

""",

"Resume_4": """

Business Analyst with Excel, reporting, and communication skills.

Some exposure to SQL and data dashboards.

"""

}

# -----------------------------

# STEP 2: Job Description

# -----------------------------

job_description = """

Looking for a Data Analyst with Python, SQL, Machine Learning,

data visualization, and statistics knowledge.

"""

# -----------------------------

# STEP 3: Keywords

# -----------------------------

keywords = ["Python", "SQL", "Machine Learning", "Pandas", "Visualization"]

# -----------------------------

# STEP 4: Keyword Score

# -----------------------------

def keyword_score(resume_text, keywords):

score = 0

for kw in keywords:

if kw.lower() in resume_text.lower():

score += 1

return score / len(keywords)

# -----------------------------

# STEP 5: Semantic Score (TF-IDF)

# -----------------------------

def semantic_score(resume_texts, job_description):

documents = [job_description] + resume_texts

vectorizer = TfidfVectorizer(stop_words='english')

tfidf_matrix = vectorizer.fit_transform(documents)

similarity = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:])

return similarity[0]

# -----------------------------

# STEP 6: Run Agent

# -----------------------------

resume_names = list(resumes.keys())

resume_texts = list(resumes.values())

sem_scores = semantic_score(resume_texts, job_description)

results = []

# Decision Threshold

threshold = 0.6

for i in range(len(resume_names)):

k_score = keyword_score(resume_texts[i], keywords)

s_score = sem_scores[i]

final_score = (0.5 * k_score) + (0.5 * s_score)

# Shortlisting Decision

decision = "Shortlisted" if final_score >= threshold else "Not Shortlisted"

results.append({

"Resume": resume_names[i],

"Keyword Score": round(k_score, 2),

"Semantic Score": round(float(s_score), 2),

"Final Score": round(float(final_score), 2),

"Decision": decision

})

# -----------------------------

# STEP 7: Display Results

# -----------------------------

df = pd.DataFrame(results)

df = df.sort_values(by="Final Score", ascending=False)

print(df)

Fraud Detection AI Agent 

import random

class FraudDetectionAgent:

def __init__(self):

self.history = []

def rule_engine(self, txn):

score = 0

if txn["amount"] > 10000:

score += 0.3

if txn["location"] not in ["India"]:

score += 0.3

if txn["device"] == "New":

score += 0.2

if txn["transactions_last_hour"] > 5:

score += 0.2

return score

def ml_score(self):

# simulate ML model

return random.uniform(0, 0.5)

def decide(self, final_score):

if final_score > 0.7:

return "🚨 BLOCK"

elif final_score > 0.4:

return "⚠️ FLAG"

else:

return "✅ APPROVE"

def run(self, txn):

rule_score = self.rule_engine(txn)

ml_score = self.ml_score()

final_score = min(rule_score + ml_score, 1)

decision = self.decide(final_score)

result = {

"Fraud Score": round(final_score, 2),

"Decision": decision

}

self.history.append(result)

return result

# Demo run

agent = FraudDetectionAgent()

transaction = {

"amount": 50000,

"location": "Russia",

"device": "New",

"transactions_last_hour": 8

}

print(agent.run(transaction))

AI agent for spam detection (Spam vs Not Spam) 

# Step 1: Import libraries

import pandas as pd

from sklearn.feature_extraction.text import TfidfVectorizer

from sklearn.linear_model import LogisticRegression

# Step 2: Create sample dataset

data = {

"text": [

"Win a free iPhone now",

"Limited time offer claim prize",

"Hello, how are you?",

"Let's schedule a meeting",

"Congratulations! You won lottery",

"Project update attached",

"Click here to get rich fast",

"Team discussion tomorrow"

],

"label": [1, 1, 0, 0, 1, 0, 1, 0] # 1 = Spam, 0 = Not Spam

}

df = pd.DataFrame(data)

# Step 3: Convert text to features

vectorizer = TfidfVectorizer()

X = vectorizer.fit_transform(df["text"])

# Step 4: Train model

model = LogisticRegression()

model.fit(X, df["label"])

# Step 5: Prediction function (Agent Brain)

def spam_agent(message):

msg_vector = vectorizer.transform([message])

prediction = model.predict(msg_vector)[0]

if prediction == 1:

return "🚫 Spam"

else:

return "✅ Not Spam"

# Step 6: Test

print(spam_agent("Win money now"))

print(spam_agent("Let's meet at 5 PM"))