Twilio has introduced Generative Custom Operators, which extract intelligence from voice-based customer interactions.
As the name suggests, the operations utilize generative AI to analyze call transcripts and enable various use cases.
Announcing the Generative Custom Operators during a YouTube video, Alex Goldman, Developer Evangelist at Twilio, explained:
It’s like having a mini data scientist living inside your call transcript.
For instance, they can score calls, track script adherence, gather competitive intelligence, and isolate upsell opportunities.
The first two use cases are a complement to contact center quality assurance (QA) programs.
Indeed, the call scoring feature automatically rates agents’ performance based on custom criteria, monitoring call quality across every customer conversation on voice.
Meanwhile, the script adherence capability is helpful in ensuring agents are following protocol.
Yet, by tracking those who are going off-script and outperforming the standard performance level, contact centers also gain insight into how they can improve their scripts.
The competitive intelligence feature is also beneficial. It notes if a competitor is mentioned and gets context on why they came up.
Contact centers can consolidate then these insights and use AI to support agents in real time, presenting live prompts to improve brand positioning during customer conversations.
Then, there’s upsell identification, with the system identifying potential sales movements in any conversation.
The feature employs lead qualification and extracts indicators of lead quality based on buying signals and sentiment.
As more contact centers focus on actively generating revenue, such use cases are increasingly coming to the fore.
Enhanced Speech-to-Text & Text-to-Speech Capabilities
Other updates from Twilio include speech-to-text and text-to-speech capabilities for the Gather and Say TwiML verbs.
This includes incorporating Google’s latest technologies, which are version 2 of their speech APIs, and DeepGram with their Nova 2 speech models.
Goldman added: “For speech-to-text, developers can now choose specifically which speech model or provider they want to use for the different parts of their application.
This choice can be made per individual prompt, which allows for greater customization based on the application’s needs.
For instance, enterprises could use one provider for general conversation and another for recognizing numbers or addresses.
The Latest on Twilio Segment
Goldman also offered updates on Twilio’s CDP, Segment, which “allows businesses to collect data from a vast number of sources, organize it into unified customer views, and act upon it to engage customers at scale better.”
Segment collects data from every touchpoint, unifying customer reactions into a complete profile.
Recent updates include the launch of Recommendation Audiences for Segment. The capability allows marketers to build audiences of customers most likely to purchase a product, category, or brand.
There’s also a Recommendation Traits tool for marketers to isolate the products a customer is most likely to buy. They may then save those items to the customer profile.
Finally, there is a Linked Events feature. It allows data engineers to define relationships between datasets for more intelligent, targeted customer campaigns.
Shortly following the release of these capabilities, Twilio announced its largest-ever customer win for Segment.
The news came as Twilio has – in early 2024 – reportedly considered selling the customer data platform (CDP).
However, CEO Khozema Shipchandler decided to retain the solution and build a better bridge to its core communications product offerings.
Join the CX Community That Values Your Voice
This is your space to speak up, connect, and grow with thousands of CX leaders. Share your voice, influence what’s next, and learn from the best in customer experience. Join the conversation today.
