The Salesforce Agent's Guide to Building an Enterprise Chatbot
Thought Leadership

The Salesforce Agent's Guide to Building an Enterprise Chatbot

Shir Fomin
March 30, 2021 by Shir Fomin

A chatbot might seem out of reach if you don’t have an IT team in place with time to code. But with a no-code enterprise chatbot platform, non-technical people like your customer support agents can build an enterprise chatbot that will not only make things easier for your customers by letting them self-serve for many issues, it will make your agents’ lives easier by automating common tasks and collecting customer information up front. Your agents are actually the best people to build your enterprise chatbot job because they’re the ones that know your customers and understand their needs and pain points best. With the right no-code platform, automating your customer experience is easy and will yield serious results for your eCommerce business, your customers, and your bottom line.

Your Salesforce agents can follow these steps to build their own enterprise chatbot.

Determine your target persona

The first step in creating an enterprise chatbot is figuring out who you’re talking to, and data is the key. Your agents can use data from your Salesforce instance to understand what your customers are trying to do and then build the chatbot to mimic key customer journeys, making it as easy as possible to get those things done. Start by building a profile for your ideal buyer, including information such as age, gender, job title, income, lifestyle, and interests. Programming your enterprise chatbot to resonate with this persona will have a big impact on your CX—for example, if your target audience is younger, you may want to use more colloquial language. But if you’re talking to experienced professionals, you’ll want to keep your content more buttoned up.

Identify your primary customer requests

Next, dive into your customer journeys and friction points to identify the main problems you need to solve. What are your most frequently asked questions? What repetitive tasks can the enterprise chatbot automate? Which issues take the most time away from your agents? What makes shoppers abandon their carts? Categorize, prioritize, and determine how your enterprise chatbot can handle these queries without escalating to an agent. 

Train your enterprise chatbot using a dataset

Once you’ve figured out what issues you want your enterprise chatbot to address, create a dataset detailing every possible way that your customers might phrase these questions so you can program a natural response. You can do this using your existing customer data or with a pre-made dataset.

Train your enterprise chatbot with available data

Dig into your Salesforce instance to gather all the information you need. You can build a database using chat and call transcripts, email replies, and form requests, as well as social media activity. This will help you detail the various ways people ask questions and use different dialects, punctuation, and so on. The more detailed your dataset, the better your enterprise chatbot will be able to recognize and discern real-world inputs. 

Train your enterprise chatbot with pre-made datasets

If you’d prefer not to rely on your own data, you can also get pre-made datasets that provide a broad range of general responses and requests, including questions and answers to obtain basic information. This can be helpful if you need to get your chatbot up and running right away and can help you train natural language processing (NLP) models faster.

Decide where your enterprise chatbot will live

Figure out which platforms will allow your enterprise chatbot to have the greatest impact. This will depend on your target audience and business goals.

Enterprise chatbots for marketing

If the primary purpose of your enterprise chatbot is generating leads, closing sales, and marketing your product, you want to make it available on the platforms where your target audience is already engaging with your eCommerce brand. Think about where your target persona spends their time—are they on Facebook, Instagram, your website? Better yet, use customer behavior data to pinpoint where they’re engaging. Put your enterprise chatbot where people will expect it and use images, videos, and GIFs as well as tone and personality to make it part of your brand. You should also make it clear that it’s a chabot so users don’t think they’re chatting with a live agent.

Enterprise chatbots for support

If your enterprise chatbot is designed for customer support, you need to put it in the places people are reaching out for help. This likely includes the page they use to process returns and refunds, your FAQ or help page, and could also include social media or messaging platforms. Do some digging and figure out where customers go when they have a problem, and then put your chatbot there to provide instant help.

Enterprise chatbots across the entire customer lifecycle

Consider everything your enterprise chatbot might offer, from initial recommendations to long-term support. Maybe it will help a customer make a purchase on your website, text them regarding order status, or provide return instructions on a messaging platform. Deploy your enterprise chatbot wherever it can accelerate resolution, answer questions, or support transactions.

Test your enterprise chatbot

Once it's programmed, you’ll need to test your enterprise chatbot for things like flow, UX, response time, accuracy, failsafes, and successful engagement. The more you test, the better it will perform.

Test within your organization

Put your team members to work running as many scenarios as they can. They know what your customers ask and can have fun trying to come up with 100 ways to phrase the same question. Just remember that they’ll be affected by their own proximity, because they probably know the answers to any question they can throw out. You’ll need someone outside of your organization who’s less familiar with your business and processes to test it, too.

Recruit your most loyal customers

The best testers are the people that will be using your enterprise chatbot. Provide an incentive for loyal customers to try to stump your chatbot and test its limits. Find out if it helps with their most pressing issues and whether or not they like its personality. 

Keep improving your enterprise chatbot after launch

Your chatbot’s life doesn’t end with its launch — it begins there. Stay on top of how it's performing, what people want, and what it needs to do to handle your latest offerings and market needs. No matter what, your chatbot needs to stay up to date with your customer needs, demands, and expectations. 

Think about new channels to launch your enterprise chatbot

Once your enterprise chatbot is thriving and your customers are happy, think about any additional channels where it can be deployed. Where are the customers you haven’t met yet? What are the newest platforms consumers are using? Where’s the next place your chatbot can go? Find out, and be there.

To learn more about how your Salesforce agents can set up and deploy an Ada chatbot for your enterprise needs, schedule a demo today.

Related posts

View All Posts