Exploring the future of customer engagement with Twilio Alpha

Exploring the future of customer engagement with Twilio Alpha

Today, we’re introducing Twilio Alpha, a new program launched by Twilio’s Emerging Technology and Innovation team (ETI) to showcase our research and give developers hands-on access to our work.

Artificial intelligence is moving at a blistering pace. Our team is launching Twilio Alpha in order to publish our early learnings as they’re happening and get developers’ hands on our work in their earliest stages – rough edges and all.

Every day, our team wakes up thinking about how the world of customer engagement is going to be disrupted as technology continues to advance, and customers’ expectations and preferences evolve. With Twilio Alpha, we will publish our work across three stages – “Pizza Box Sketch”, Prototype, and Developer Preview, to give customers and developers a firsthand look into our research and let them shape our work.

Twilio Alpha’s first research area: the impact of autonomous agents in customer engagement

Our first research area is focused on autonomous agents and Artificial General Intelligence (AGI), which we believe will create fundamental disruption in our industry and to customer engagement as a whole. While there are a multitude of ways that these developments can change the status quo, our team is particularly interested in exploring the impact on the three foundational pillars of Twilio: Customer Data, Communication Channels, and APIs. Below, we outline some of the foundational questions that will drive our research.

The future of channels in the world of AGI

The way communications are delivered to consumers will fundamentally shift with AGI.

As artificial intelligence improves and the quality of channels like voice improves, customers will interface with personalized autonomous agents acting on behalf of brands without any human intervention, fundamentally changing the way that businesses manage their communication channels. Conversational interfaces like chat and messaging will improve with LLM advancements – a shift to autonomous agents might reduce the need for companies to have their own channels of communication, and instead maintain interfaces that AI can interact with.

We are looking to understand how digital communications will evolve in the world of AGI, asking the question: what role do traditional communications channels play in a world of AGI and how will they evolve?

The role of customer data & personalization

A fundamental part of unlocking exceptional customer engagement with autonomous agents will be giving AI the most complete understanding of a customer possible. While the rapid growth of context window sizes (essentially the working memory of an LLM) enables companies to pass more data about a customer to LLMs, the best way to achieve customer awareness and memory remains largely unsolved today.

We believe that in order for AI to unlock a complete customer picture, it needs to effectively learn and reference interactions across the whole customer journey, including a customer’s preferences and historical interactions with the company.

At Segment today, we aim to provide companies with a complete picture of their customers across various systems. Now we’re also asking: in a world where autonomous agents dominate customer engagement - what customer data is needed to support an immediate and accurate interaction, every time?

The importance of APIs and Developer Experience with AGI

An aspect that makes autonomous agents fundamentally different from previous technologies is their ability to interact with the world to perform tasks.

Today, an agent’s interface with the world is largely limited to the functions & APIs that humans give them access to, thereby limiting their capabilities. However, more recently we've seen new systems that begin to remove these limitations and dependencies. Whether it’s Cognition Labs' offering "Devin" that can write its own code, use the command line, browse the web, learn from blog posts, and debug issues, or Rabbit's Large Action Model (LAM) that is centered around performing actions in the browser, we’re seeing some of the dependencies for dedicated APIs go away.

APIs have long been the foundation of communication between systems in general, and have been the very foundation of Twilio's success. With AI being able to interface with a multitude of systems and write its own code in the future, we’re asking: what do the optimal interfaces and developer experience (DX) look like? Will we have to rethink how we build APIs in the world of autonomous agents?

Twilio AI Assistants launches to Developer Preview today

Today we are announcing AI Assistants to Developer Preview as Twilio Alpha’s first step in our exploration of the future of autonomous agents in customer engagement.

Twilio AI Assistants is a no-code builder for Conversational AI use cases built on top of CustomerAI, with tight integrations to Twilio channels and Segment. Using the power of Large Language Models, AI Assistants are capable of solving problems automatically using tools (APIs), knowledge (websites, documents, etc.), and customer awareness (CustomerAI); keeping the system in check with Guardrails; and seamless hand over to other systems or humans as needed.

We believe developers will shape the future of customer engagement, so we want to get our work in their hands early and will keep shipping as our research progresses.

What’s next

We are excited to continue our exploration around the future of customer engagement against the backdrop of unprecedented technological advancements. You can follow along on twilioalpha.com, and join the waitlist for Twilio AI Assistants. We can’t wait to see what you build!

Kat McCormick Sweeney leads business development for the ETI team. She gets pumped about solving customer problems with her team. When she’s not thinking about customer engagement (it happens), she’s exploring the outdoors with her family, reading fiction or watching the Food Network. Connect with Kat at kmsweeney [at] twilio.com.