We’ve noticed that there aren’t many articles that describe how the process of implementing AI voice assistants looks like in detail. We decided to start a series that will dispel any doubts about the subject and prove that any company that struggles with repetitive customer interactions can benefit from adding a virtual agent to their team. Lead generation automation is a great option for companies that are afraid of using AI agents and want to start small before going all in. We get it, multiple bad experiences with automated voice services make most people queasy. It’s is a good way to see the benefits of voice agents: money saved, increased effectiveness, improved customer experience, even increased team morale, without making a big investment. But enough marketing, let’s dive in.
Let’s start with problems that virtual agents can help you solve:
The key to a successful implementation of an AI virtual agent is finding the use case that will deliver ROI. This leads us to the question of how much of your sales process you can automate.
AI agents take care of repetitive tasks, those ones that sometimes make people feel like machines. They do it more effectively and cost less. But you can’t contain everything with automation.
Virtual agents don’t have the judgment and empathy that are required to run successful sales negotiations. They support your existing team and empower them to up their skills, so your sales representatives can handle conversations that only a human should handle.
Now, that you know what virtual agents can help you with and what should be left for your human team, let’s look at how lead generation AI implementation looks in real life:
As mentioned before a key step in creating a successful virtual agent is finding out what process will benefit your business when automated. Our team has many successful implementations under its belt and is happy to help find the best solution for you.
One of the lead generation processes that virtual agents are great at is cold calling. Repetitive and frustrating for your sales team and costly for your business. AI agents don’t have a problem with hearing “no” over and over again, and when they hear a yes, they can route the call to a live agent. We’ll use cold calling automation as an example for the rest of the article.
When cold calling, you only have a few seconds to grab someone’s attention. It’s important to figure out how to make the message more personal or mention a competitive advantage and lead with it. Just like a live agent would do.
Let’s say work at an insurance company and want the virtual agent to call all customers who are close to the end of their car insurance contract. Here’s how this conversation could go:
“Hello, I’m a virtual agent calling from insurance company X, I’ve noticed that your car insurance ends soon, would you like to extend it?”
If they say yes, the virtual agent can route the call to a consultant or ask when the consultant should call back.
If they are not sure, the virtual agent can address certain concerns and answer additional questions.
If they say no, the call ends politely.
Now that the script is ready, your virtual agent can be configured in the Talkie app. It’s quite easy, you click and point and drag and drop. No coding required. Our system offers many predefined components that make the process quick. As you’re new to the app our team helps set it up and shows you all the nooks and crannies, so you feel confident when you want to make changes later on.
Once the virtual agent is created you can start having conversations with it. The voice is synthesized for testing but you can add your own recordings to make the conversation more natural. We actually recommend it, as it increases the conversion rate. For example, one of the live agents on your team can lend their voice to the virtual agent.
On the one hand, it’s quite fun to talk to the virtual agent, on the other, it’s an opportunity to see if everything works correctly. For example, you can check if the agent understands different ways of giving the same answer. You need to train it as you would with a live agent. The goal is to make the user experience as good or better than with a live agent. Talkie team also helps with training, so there’s always a helping hand if you need it.
When training your agent, you should also figure out what the person on the other end might say that is not a part of the script.
Sometimes when a virtual agent calls to offer one product, the customer asks about something else in your offer, e.g., bike instead of car insurance. AI agents can be trained to listen to those keywords and address questions containing them gracefully, so the customer hears: “Oh, so you’re interested in bicycle insurance, we offer it too, I can connect you to one of our specialists… “ instead of the old “I don’t understand, could you please repeat?”.
Virtual agents are there to support humans, that’s why it’s important to make sure that every single person on your sales team knows how the virtual agent works, how it routes calls to them or saves dates to call the prospect back. If your team uses a CRM, it might be worthwhile to connect it with Talkie so all customer information is in one place. This way customers get smooth experience and you make the most out of the virtual agent.
Once we’re sure the AI agent is ready, it’s time to go live. All conversations are saved in the app. You can go through transcripts and listen to all conversations, so you stay in the loop and can dynamically train the virtual agent. But as lead generation scripts are often quite straightforward, virtual agents generally don’t require much “technical” training after they go live. At this stage, it’s more about finding the script that provides the best conversion rates.
Talkie has robust reporting capabilities, so you can see what makes the prospects say yes and what makes them hang up. You can also easily run A/B tests to compare how a different message or time of day can change customer behavior. What’s great about virtual agents A/B testing is that you control all the conditions, for example, 200 people can be called at the same time, with live agents it’s simply not feasible.
Once you’ve set up your first virtual agent, the time needed to create another is cut in half. This way you can gradually go beyond lead generation and automate more and more processes in your company.
All your virtual agents can share data, so if one is trained, others learn too which makes it easier to manage them later on.
The whole process takes from 5 to 10 days and then you can call your prospects whenever it suits you. As you can see, even though underneath it all, there are complex artificial intelligence algorithms that let virtual agents mimic natural conversation, you don’t need to deal with the technical side.
Want to try automation with Talkie? Sign up for a demo on our website. Working in a different industry? Check out these virtual agent use cases:
Are you looking for solutions to automate your sales processes?