Crafting the perfect transactional AI Script

Today's call center landscape is undergoing massive development, with the onset of AI-driven calls. While being at the forefront of the race to provide the best AI solutions for today's call centers, Callchimp.ai also brings forward an upgrade to the existing infrastructures for faster adaptability.

Have you ever picked up your phone to hear "Hello Gourav, this is to confirm your transaction of Rs. 81,000 towards ABC Pvt. Ltd. Do you confirm the payment, or would like to register a complaint?". Calls like these almost have the same script except for some particular values in this case the customer name and amount of money. These are transactional calls. While transactional calls like these served almost static content, Callchimp.ai brings an exciting upgrade to this.

Apart from serving static messages with variables, Callchimp's transactional calls are now fully powered by AI to have a proper conversation, process responses, and mark dispositions. This cutting-edge technology empowers businesses to engage with customers more naturally and efficiently, driving improved outcomes.

Writing a transactional AI call script

Writing a transactional AI call involves all the best practices needed to write an AI script. To get a better understanding of how to write the perfect AI script, please refer to our blog on "Creating the perfect Callchimp campaign".

Transactional calls use Callchimp's Dial AI campaign and, require certain placeholders that Callchimp's AI can replace with original values to place a perfect call.

How to declare placeholders for a script.

Let's say we have a script that places a call to the customer to confirm the purchase of a product, it would be something like this:

Hello Gourav, this is a call from ChimpMart to confirm your purchase for Callchimp Pro , costing Rupees Forty thousand. The order was placed at 24 April 2024. Can you please confirm if you'd like to continue with the order.

In the above script understanding what the possible placeholders are a key strategy to writing the script. In this example, the placeholders are :

• Gourav: first_name

• Callchimp Pro: product_name

• Forty Thousand: product_price

• 24 April 2024: order_date

To include placeholders into the script, we wrap them around curly brackets to denote a placeholder, so Gourav becomes {first_name}.

Hello {first_name}, this is a call from ChimpMart to confirm your purchase for {product_name} , costing Rupees {product_price}. The order was placed at {order_date}. Can you please confirm if you'd like to continue with the order?

Thus you can convert your static script into a transactional script.

Making better transactional scripts

Using proper string templating

While writing a transactional AI script, providing some context and writing all the dialogue in quotes is always better.

Greet the user by saying "Hello {first_name}, this is a call from ChimpMart to confirm your purchase for {product_name} , costing Rupees {product_price}. The order was placed at {order_date}. Can you please confirm if you'd like to continue with the order?"

Using SSML

Speech Synthesis Markup Language (SSML) is the best way to customize how the call sounds and is played allowing more control over audio. Integrating the above script with SSML,

Greet the user by saying "<speak>Hello {first_name}, this is a call from ChimpMart to confirm your purchase for {product_name} , costing Rupees {product_price}. The order was placed at {order_date}.<break time="750ms" /> Can you please confirm if you'd like to continue with the order?</speak>"

Integrating AI scripts

Callchimp's DialAI campaigns add the next level of interaction to transactional campaigns. After the transactional part of the script, simply add proper prompts to guide the customer through the AI call. For the script above, we can add a few more lines.

Greet the user by saying "<speak>Hello {first_name}, this is a call from ChimpMart to confirm your purchase for {product_name} , costing Rupees {product_price}. The order was placed at {order_date}.<break time="750ms" /> Can you please confirm if you'd like to continue with the order?</speak>"

If the user confirms or responds positively, say "Thank you for shopping with us! We're processing your order. Goodbye."

If the user responses negatively, say , "Thank you for confirming, we are canceling the purchase. We would love to have you shop with us. Goodbye"

Repeat the greeting only if the user asks or is unable to understand.

And that's it

Callchimp's innovative new transactional AI calls lay the foundation for businesses to scale their growth exponentially! With the onset of this technology, we not only empower businesses to place calls to customers but also to collect information and make strategic decisions, thus scaling their ideas to their fullest potential.