Strategies for Implementing AI in Customer Service

One of my favorite exercises as a Customer Service leader has been uncover the drivers of support friction and support volume and then come up with a strategy to address the primary drivers. Generative AI has created new opportunities to automate support as well as make it easier.

I recently partnered Miso Robotics' Product team on their AI strategy. They were overwhelmed with all the ways AI could be integrated into their organization. I provided an overview of how AI is being use in Customer Service, which tools professionals in the space recommend, and an industry overview. That’s great - but until you have a strategy to focus on, which tool to implement doesn’t matter. Below are three strategies - which one you choose can and should change as the needs of your organization change.

But first..

Before you figure out a strategy do an assessment. It's crucial to understand what you know, the strengths and limitations of the support team and broader organization, the data and actions your team needs to support customers, and even the sentiment around AI in your team.

If you don’t have data, you need to get it.

If your data is shit, you need to clean it.

If your team is worried they’ll be replaced by AI, your strategy should involve empathy and building trust.

etc. etc.

Understanding your customers' pain points and identifying areas of friction for your support team is paramount. If this information isn't readily available, measuring and gathering feedback from both customers and your internal teams is a solid starting point. This leads us to our first strategy: The Voice of the Customer

Strategy 1: The Voice of the Customer

Foundational metrics like Average Hold Time (AHT), resolution rate, and Customer Satisfaction (CSAT) are valuable but they provide more of a yardstick then the narrative. Understanding why customers are contacting you and how complex these issues are for your team to resolve is imperative. Otherwise you may invest in a solution like deploying a Conversational AI agent only to find it’s unable to solve a meaningful amount of your issues.

This strategy is ideal for organizations that are:

• Operating without clear customer insights
• Transitioning to a customer-centric model
• Aiming to understand customer feedback before allocating resources to CS infrastructure, product quality, operations, or functionality

Implementation ideas:

1. Utilize platforms like Monterey.AI to triage customer feedback, identify themes, and analyze your data using natural-language queries.
2. Leverage AI tools like Claude.ai to process and analyze customer survey results, helping you pinpoint key areas for investment.

Strategy 2: The Make-it-easier Approach

This strategy aims to shift issue resolution from highly technical teams to less specialized ones, thereby simplifying support processes.

It's particularly effective for teams that:

• Frequently rewrite SQL queries or maintain lists of troubleshooting notebooks
• Handle complex tickets typically resolved over hours or days
• Work with intricate systems

Implementation ideas:

1. Equip your agents with a "Super Tool" like dataland.io, which integrates various data sources and APIs, allowing problems to be solved from a single interface.
2. Develop APIs for common support actions, enabling workflow automation and reducing effort for complex issues.
3. For teams already using APIs, consider implementing an AI-native workflow automation platform like n8n to streamline processes that currently require manual intervention.

Strategy 3: The Make-it-go-away Method

Support issues, not your customers.

This two-pronged approach focuses on:

1. Maximizing support automation (with some teams achieving up to 92% automation)
2. Addressing root causes of support issues, whether they stem from product confusion or functionality gaps

This strategy is ideal for:

• Teams handling high contact volumes
• Organizations dealing with a large number of easily answerable questions or tasks

Implementation ideas:

1. Deploy conversational AI agents across chat, email, and phone channels. Ada comes highly recommended, but ensure thorough knowledge base development, testing, and experimentation before a full rollout.
2. Invest in robust solutions like Monterey.AI to identify and rectify the primary drivers of support volume, focusing on features or issues that generate the most tickets.

By adopting these strategies and leveraging AI technologies, organizations can significantly enhance their customer service operations, reduce support complexity, and gain valuable insights into customer needs and pain points. Remember, the key to success lies in choosing the right approach for your specific situation and implementing it thoughtfully and systematically.

Want more details?

Hire me to come and work with your team, talk to your organization or… be patient, more resources are coming.