Introducing Radial

Customer service is broken both for frustrated customers, and the support teams trying to help them.

Support teams try to keep up, but the traditional model of hiring and training more agents doesn’t scale. AI was supposed to solve this problem, but first-generation tools stopped at FAQ chatbots and basic tier-1 automation.

But customers expect instant, accurate, empathetic support, 24/7. In a competitive world, you either deliver great, personalized customer experience, or watch your customers churn and move to your competitors.

Today, we’re excited to announce Radial, a new support platform built from scratch to bridge this gap between customers and support teams.

Play

Deflection isn’t resolution

Most AI support platforms optimize for deflection rate: the percentage of conversations handled without human involvement.

Deflection doesn’t automatically mean “resolved”. In many cases, chatbots optimized for deflection push customers through frustrating flows and link out to generic FAQs that don’t actually solve their problems.

This leads to even more frustrated customers escalating to your human agents, or worse, customers that abandon your product and become detractors for your brand.

Deflection metrics might look good in dashboards, but deflection is not resolution. Customers know the difference, and they remember it.

Getting from deflection to resolution

A resolved issue is one where the customer got what they needed and aren’t facing problems anymore. For example:

• Customer reported an API failure → failure triaged and corrupted data fixed
• Package lost during shipping → replacement dispatched
• Customer accidentally overcharged → investigated and refund issued (as per company policy)

Getting an issue to a good resolution is often a research and investigation problem. It requires:

• Understanding intent - not just matching keywords or searching FAQs, but often conversing with the customer to figure out what they actually need
• Access to real data - customer history, account state, transaction records
• The ability to act - not just recommend knowledge, but actually execute actions: issue refunds, reroute shipments, update subscriptions, manage quotas
• Good judgment - knowing when to handle something autonomously and when to escalate

This is what Radial is built to do.

How Radial works

Radial’s AI agents are deeply integrated into your business and pull context from everywhere. They connect to the systems you already use: from CRMs, to order management systems, billing systems, communications systems, and even your custom software. They’re also trained on your policies, your brand, and your tone.

When a customer reaches out, Radial doesn’t just search for a matching FAQ. It pulls the relevant data, evaluates the customer’s situation against account state, company policies, and prior interactions, then takes action in seconds.

For example, if a customer asks about a missing package, Radial can check the order status in the fulfillment system, check shipping confirmation, and even initiate a replacement shipment while adhering to internal company policies.

Radial agents are also omnichannel, available across in-app chat, email, SMS, and voice to meet your customers wherever they are.

Radial is enterprise-ready, with configurable approval flows, escalation policies, audit logs, and action limits so companies retain control over sensitive operations.

Customer experience is a competitive advantage

Prior to Radial, Karan and I founded Warrant, a Y Combinator backed software infrastructure startup.

Warrant’s core product was an API used by companies to add authorization and permissions to their software applications. As a small team, one of our biggest competitive advantages was the high-touch onboarding and support we personally provided. It helped us win customers, but it became difficult to scale as the company grew.

Radial is the AI-native support platform we wish had existed back then. It autonomously handles a large portion of support interactions end-to-end. For issues requiring human involvement, it performs the majority of the investigation and hands off full context to the support team so that a human can take it across the finish line.

What’s next

Radial is now in beta. Over the past couple of months, we’ve been working closely with early design partners to deploy Radial within their support teams.

If you’re looking to supercharge your customer experience, chat with us.

Or chat with Radial - click on the icon below on the bottom-right.