AI and Automation
Beyond the chatbot: Introducing targeted automations powered by WFM insights

Beyond the chatbot: Introducing targeted automations powered by WFM insights

JUMP TO SECTION

In the three months since we released Assist, our issue resolution engine and agent copilot, we’ve been working hand-in-hand with our customers to further unlock the power of AI for support. Today, we’re excited to announce our latest Assist product release: automations that use your workforce data to power targeted use cases that match your business needs. Assembled Assist customers can build powerful automation flows with the help of a dynamic staffing timeline visualization to guide the strategic deployment of AI. We’re helping companies go beyond the basic chatbot and use AI in a way that matches their business needs—all while giving customers the top-tier experiences they deserve.

Assembled has always built product with the goal of increasing efficiency in support operations. We started with workforce management (WFM) with the aim of ensuring every team has the right person in the right place at the right time to address customer inquiries. And with AI, we’re now making sure that the right answer is provided during those interactions.

Don’t make customers feel deflected

Let’s dig into that last word: interactions. Each touchpoint with a customer should feel helpful, personal, and valuable. Almost all companies have some form of self-service, maybe with a chatbot to help direct folks to the right resources. The term typically used to describe this is “deflection”—a word that implies the process is preventing a customer from engaging with someone at the company. But a deflection shouldn’t feel like a deflection to the customer: it should just feel like a speedier path to a successful resolution. All too often, this just isn’t the case: chatbot flows can feel robotic, impersonal, and frustrating.

That’s because too many chatbot flows are created without a transparent view to what’s actually happening with data. The data within your WFM—including how long it takes agents to do certain tasks—should inform your approach to AI, not the other way around.

That’s why we built Assist to be an AI-powered issue resolution engine—not just a deflection tool. That resolution might come from an automation or it might come from a human agent, but our goal is to make sure it’ll never feel like a deflection.

Targeted automation for personalized service

The answer of where to focus on automation lies within the data you already have on hand. Assembled Assist visualizes your staffing timeline and allows you to see the predicted impact of different types of automation. With this dynamic visualization, you’ll be able to see the best opportunities to automate while understanding the impact on your staffing so you can optimize both agent and AI resources.

Once you’ve identified the opportunity, you’ll be able to configure a resolution path to guide the steps needed to resolve the ticket. But make no mistake: this is not a rigid chatbot flow that you’ll have to constantly revise. Using generative AI, the automations you set up will produce tailored responses that connect to your existing documentation and brand style guide. That means replies feel natural and engaged, never like the person is being blocked from the help they need.

Take Honeylove for example: they’re using Assist to fully automate returns via email, including sending customers a unique link to download forms and labels. And yes, Assist remembers to include their signature purple heart emoji. That’s a big difference from a chatbot, which would be able to give a customer a link to help center documentation where they might be able to self-serve into the return process.

And Assist automations are highly flexible: toggle automations on and off, add an auto delay on sending so messages aren’t too bot-like, and even select the volume of automated responses that are sent. For example, you can adjust automations to step in when volume gets wild. Every Assist automation has a “confidence level”—basically, a percentage of how accurate the AI grades its own response based on the materials it has at hand and the contents of the interaction with a customer. When volume is low, you might have automations set so the confidence level on a reply is extremely high before anything gets automatically sent to a customer. But when things get hectic and you’re about to break SLA, you can configure Assist so it can come to the rescue and automate a larger chunk of interactions.

WFM + AI: A virtuous cycle

We’re just getting started with Assist! We’re building towards a future where the combination of WFM and AI helps you create a virtuous cycle within your support function:

  • Reveal the right opportunities for automation through analysis of your tickets
  • Deploy smart automations that automatically adjust with the ebb and flow of your business
  • Improve your staffing and forecasting based on what you learn through the process

Automation is amazing, but the real magic happens with the connection to WFM. Add it all up and you’ll create a continuous process that unlocks the full potential of AI for support.