Updated by Amanda Porter
Thank you for considering Capacity. We’re thrilled for the opportunity to begin working together to automate support for your employees and customers. If you’re curious about our process or how quickly you can start realizing the value from our product, this guide will answer questions and provide the details you need to start doing your best work!
Our implementation process actually starts during the sales cycle. Our team of experts will work with you to define your requirements and scope your project early in the sales process. This allows you to get up and running in as quickly as 30 to 45 days of signing your service order.
How we do it:
- We meet regularly to set expectations and to keep the project on track.
- We prep your team to submit the necessary knowledge for your use case and integrations.
- Our expert Customer Support team will quickly and efficiently load your information into the knowledge base to train Capacity.
- Our Integration Specialist will connect your apps to Capacity (if applicable).
- Your team will train on the system with guidance from your Customer Success Manager.
- We will test everything with your implementation team.
- We will release Capacity to the designated end users.
We know that AI training isn’t something that many of our customers are prepared for. And we also understand that every rollout is different for every customer. We created this ebook to answer frequently asked questions so everyone is on the same page from day one.
Meeting to Set Expectations.
The first step that we always take includes getting acquainted with our new customer, reviewing the kick-off plan, and determining what our customer wants their knowledge base to look like (we’ll get into this in the next section).
Your Capacity salesperson will introduce you to your Customer Success Manager, Director of Customer Success, a member of our Customer Support team, and the Solutions Engineering Lead for integrations (if applicable). Prior to the initial call, your salesperson will make sure that everyone on the Capacity team is fully aware of your goals from the beginning, so we can efficiently set your team up for success.
Everyone on your team will receive a comprehensive look at Capacity’s features and capabilities, and learn how Capacity will empower your team to do its best work. Essentially the more knowledge that Capacity understands, the less intervention is required from your team.
From here, our team will walk you through the specific steps and expected outcomes based on your contract. We encourage our customers to come to this meeting with an understanding of the content that they want to add to Capacity, as this will jump-start the process and help us map out subsequent calls. Not sure where to begin? Just ask us, and we’ll be happy to share best practices from other customers.
Building the Knowledge Base.
To expedite the implementation process, it’s critical to provide us with information as quickly as possible. It should also be noted that the more information you want Capacity to learn, the more time it will take to train and test. Depending on the contract type, we recommend customers start with 250 frequently asked questions for the initial set up to make sure everything is up-and-running within the timeline. You can always add more questions later.
We understand that this is a new concept and that it isn't common to have a clean repository of knowledge easily accessible to train the AI. So we’re sharing a list of helpful information that can get you started.
Examples of content for an external use case (customer focused):
- Company-specific or industry-specific acronyms. *Capacity will add our current acronym packs for customers in industries that we’re frequently involved with.
- Frequently asked questions and the corresponding answer.
- Email threads from your customer support inbox.
- Information that your subject matter experts have accumulated throughout the years.
- Source documentation.
Examples of content for an internal use case (employee focused):
- Company-specific or industry-specific acronyms.*Capacity will add our current acronym packs for customers in industries that we’re frequently involved with.
- Frequently asked questions and the corresponding answer.
- Employee handbooks.
- Company policies and benefits.
- Email threads from your internal helpdesk ticketing system.
If your team lacks any of this documentation or information, we always recommend our customers work with their teams to get insight into the type of questions that you receive and the respective answers.
Curate the answers from your organizations’ subject matter experts. Tapping into their knowledge is the best way to move the needle and be as efficient as possible. These SMEs could be in the IT and HR department if you're focused on supporting your internal team or your customer support team if you’re focused on supporting your customers.
Integrating with apps is more common for customers with an internally-focused use case. If you want Capacity to connect with any apps, it’s important that you share information to get your integrations set up along with the information that you want to be added to your knowledge base. This enables our App Integration Specialist to secure proper connection prior to testing and training.
We’ve found that the following list of information can help streamline the app integration process:
- A list of the apps that you want Capacity to integrate with.
- The links to API documentation and/or PostMan collections.
- Access and credentials for OAuthorization and/or OAuthentication.
- The list of questions that your users can ask Capacity to access information in your apps of choice (we refer to these as app skills).
Training Your CoPilots.
As you gather information for the knowledge base and app integrations, we’ll start training your team’s CoPilots. These are the individuals who will be dedicated to managing your knowledge and mapping new questions. During training, all CoPilots will have secure access to the back-end of Capacity called the CoPilot Console. We’ll provide instructor-led training to show your CoPilots how to load information into the knowledge base, create folders, manage the Helpdesk, find and use our Analytics, and update settings.
During training, our Customer Support team will provide best practices on how to add new information to the knowledge base. As a general rule of thumb, we typically recommend that questions and answers are added in a very short and sweet manner like a text message. Where long responses are required, we’ll show you how to use hyperlinks to websites and documents stored in the Capacity Cloud Drive. After the live training session, we give the CoPilots a week to digest and poke around the system. Once they dabble around in the tool, we host a second session for questions.
When choosing a CoPilot, It's important to make sure they are subject matter experts on the topics that you plan to host and share via Capacity, so they can efficiently add in the correct information.
How many CoPilots do we need?
This depends on your company’s preferences and size. We’ve seen companies succeed with a dedicated CoPilot for each area of knowledge and we’ve also seen companies succeed with a pair of CoPilots who are responsible for managing the entire system.
Why is a CoPilot important?
CoPilots are responsible for monitoring incoming questions, so they will be aware of the trending information. CoPilots will know exactly what information needs to be added or updated in the knowledge base. Also, Capacity can be the first to know if a call center or website is down because customers will send messages through the bot asking questions, so it’s a good idea to have someone monitoring the inquiries as they come in.
How long should they spend time on the console?
Typically, if you have a good knowledge base that’s seeded with the correct responses and exchanges, your CoPilots won’t need to do a ton of triaging. At a basic level of usage, we recommend a CoPilot checks the console once or twice a day. But if you’re using it as a customer support tool, then we recommend having a dedicated CoPilot in the system throughout the entire day.
Adding and Storing Knowledge.
Once you deliver all of the information outlined in the “Building Your Knowledge Base” section, our Customer Support team will start seeding the information into the knowledge base in the form of variants. Variants are all of the different ways to phrase a question to receive the same answer.For example, the following variants could be asked to access the number of PTO days each employee is granted:
- “How many PTO days do we get?”
- “How many vacation days do we get in a year?”
- “PTO benefits.”
Our Customer Support team recommends thinking outside of the box and putting yourself in the shoes of your users to provide the most effective variants.
What takes the most time?
The more variants, the better the match rate. Even after providing the template of the FAQs, our Customer Support team will help add in a handful of potential variants, so Capacity has a high match rate.
We’ve found that the shorter the response, the better it is. There is a 3,000 character limit per response in Capacity to make sure user experience is optimal for a chatbot. Regardless of your use case, users will always want a simple and straightforward response. If you share a response to a question that’s really long, our Customer Support team will have to go in and add in a hyperlink or summarize the answer to our best ability.
As a rule of thumb, it could take up to 3 to 4 minutes to add each exchange, including adding the different variants and summarizing each response.
Integrating With Apps.
Capacity makes it easy to share knowledge and automate workflows by integrating with your most important day-to-day tools. Our App Integration Specialist will set up multiple endpoints per app in our Developer Platform, use those endpoints to build out the guided conversations (based on the questions that you provided), and test it to make sure that Capacity is providing the data from the endpoints that were originally set up.
Testing and Training.
We ask our customers to gather a dedicated group of users to test and train the AI before rolling it out. This includes asking as many questions as possible and seeing what information Capacity will return as an answer. The testing process is simple: If the answer is good, give it a thumbs up. If the answer is bad, give it a thumbs down. It is important to provide additional context if the answer is bad. This feedback will come through as a ticket to the backend where our Customer Support team and your CoPilots can improve the answer for future use. Our Customer Support team will also analyze the results from the test and share any trends that might have become apparent.
The teams that we’ve seen have the most success going live to production in a quick timeline are those with a solid core group of CoPilots who are dedicated to the initiative of working in the console. It’s important to have team members who are willing to put in the time to stringently test by repeating the process over a few weeks. Rather than testing and updating the knowledge base once, completing this process a few times is what leads to a match rate of 85%-95% instead of 65-75%.
Encouraging this type of behavior from the beginning is important to the big picture.
Ongoing Support After Rollout.
After the implementation, our Customer Support team will continue to support your team by monitoring and triaging any open tickets when applicable. However, if there are any compliance requirements that our customer support team isn't equipped to help with, they will refrain from assisting with your knowledge base.
We’re here to offer support when looking for bugs, clearing any gibberish and spam that might come in, and helping improve the helpdesk based on the types and frequency of questions that come in. If for some reason the questions coming in after implementation don’t match what you originally thought users would ask, we’ll step in to help analyze and improve that content as well.
If you add a new person to your team or transition responsibilities, we’re more than happy to help train them on the platform. And, as the platform evolves and new features are rolled out, we’ll also offer training to get your team acquainted with how it works.
After rollouts, we’ve noticed that our customers see stellar results as long as they encourage employees and customers to use Capacity. Because your team and customers don’t know much about it, and it’s a new technology, people may be confused and reluctant to use it.
If you’re rolling it out to your customers, make sure that your Customer Support team is bringing it up to customers as they call in. Create an expectation that they suggest the use of Capacity when customers call in that reaffirms users where they can access it. Make customers aware that Capacity is another avenue to self serve and quickly get answers to their questions.
Try something along the lines of:
“To avoid long wait times, you can go to our website and ask our digital assistant, Capacity, to help you with some of your basic questions.”
In an internal use case, if an employee is going to HR to ask a question instead of Capacity, make sure your HR team is reminding team members that Capacity is available to help.
After answering a question, they can always ask:
“Did you know that you can always ask Capacity these types of questions to get immediate responses?”
After Capacity goes live, it’s important to drive users to that channel for it to be successful. But before Capacity is live, you can also drive adoption with creative ways to make people aware that this tool is coming. Consider sending sneak peek eblasts to your customers and/or employees to generate excitement. Our customer, AmeriSave, held a contest of naming the bot so everyone was involved and aware of Capacity before it was implemented.
Before we get started, we'll work with your team to define your guidelines for success. We will make sure to share this information with your team during our recurring meetings.
If you’re wondering how Capacity is performing, you don’t need to wait for a scheduled meeting with the Capacity team. It’s very easy to see this information at any time through the Analytics Dashboard.
At Capacity, the metrics that we consider successful include:
- A minimum match rate of 80%.
- Average response times under 3.5 seconds.
- More than 4 variants per question.