How B2B Companies Can Succeed With AI In Customer Support

Let's get this straight first: Customer support is especially crucial in B2B SaaS companies because it directly ties to revenue. Period. Great support strengthens relationships, boosts retention, increases upsells, and extends customer lifetimes. Besides, happy customers are the best advocates; they are the most effective growth funnel. (Personally, seeing customers love what we're building and get real value from it makes all the all-nighters worthwhile. It gives me the ultimate dopamine hit 🥹.)

What’s Unique About Customer Support in B2B SaaS Companies

An emphasis on technical support

In B2B SaaS, we are not just dealing with simple queries like “how do I reset my password”. Issues are often complex, involving customer-specific concerns, edge cases, and even system bugs.

This means a significant portion of issues require technical investigation and thus there’s a bigger emphasis on technical support. Tech support engineers need to dig into logs, dashboards, and customer configs to troubleshoot and triage to engineers.

Tech support is not an easy job; it requires understanding the ins and outs of the product, its technical specifics, and a lot of engineering training. When I worked as an engineer at LinkedIn on the recruiter tools, I got ticket escalations a lot. Juggling between 3 different logging systems, dashboards, and databases to debug why a company couldn’t post jobs wasn't fun at all.

Thought: how to provide tech support engineers with investigation tools to help them troubleshoot faster?

Multiple teams involved in solving an issue

The complexity of these issues often requires coordination across multiple teams, like customer support, engineering, and product.

One big reason is that the support team needs to constantly stay updated with product changes. While product release notes and triage meetings help, there can still be misses, especially with edge cases or customer-specific functionalities. So, the support team often has to ask engineering and product teams to get the latest product knowledge to diagnose and resolve. Relying on internal teams for knowledge makes solving issues feel like a waiting game.

Thought: How to give support team fast access to the up-to-date product changes?

Human touch is crucial

Since each customer’s case is different & more complex, and because each customer has a larger contract size, personalized support is highly valued and also helps foster long-term relationships. Self-served support is simply less attainable and thus less focused on.

As a result, enablement of support team is a major focus - e.g. providing them with the information and assistance needed to resolve issues quickly and accurately. We should find opportunities to use automation that don't interrupt the natural flow of human communication while enhancing efficiency.

Thought: how to use automation that don’t compromise the human touch?

Support requests are closely tied to product prioritization

If a customer has a contract size of $100k, you would want to listen carefully to what they are asking for or need help with, and their questions and requests will likely influence product feature prioritization.

Support requests provide direct insight into what customers need, and thus companies wants to generate insights from all support requests to have a wholistic picture on customers’ needs (and their contract value) to determine which features to prioritize.

Thought: how to obtain real-time insights from support requests to understand what customers want?

Slack as support channels

Many fast growing B2B companies have dedicated Slack channels for each customer, but this setup makes tracking requests difficult. When they have 100 or more customers, relying on Slack will become impractical and context switching will be a nightmare. There need to be a more robust system to manage and track customer interactions.

Thought: how to aggregate and better track customer requests from Slack channels?

Focus on getting the right answer rather than a fast one

This might be slightly controversial, but I’ve learned that many B2B customers care more about getting the right answer in one go rather than merely a fast response. There are often a lot of back-and-forth in B2B customer support, so customers prefer to get their problem solved in one interaction, regardless of the amount of time spent.

Thought: how to gather customer context effectively to minimize the back-and-forth?

What Tools Can Solve These Challenges?

Luckily, we are entering a new AI age and the challenges are being solved!

Duckie - AI support assistant

We built a full-suite AI support assistant for B2B SaaS companies. We focus on empowerment, giving support teams the tools they need to resolve tickets quickly and accurately while maintaining the human touch for customers.

Duckie supports all tiers, from T1 customer support to T2 technical support, and also bridges communication between support and engineering.

Key Features:

• Technical support assistance - log investigation

Duckie diagnoses technical issues using logs, code, dashboards, and customer data, to help technical support team troubleshoot and triage tickets quickly.

‍

• Knowledge search
  

Duckie scans knowledge sources, such as documentations, past tickets, Slack history, to instantly find up-to-date and accurate product knowledge useful for resolving tickets.

• Automated engineering updates

Duckie keeps the support team informed by linking JIRA tickets and pull requests to support issues, automatically posting updates. It also tracks code commits to report on product behavior changes, so the support team can always access the SoT of product changes.

Omni channel tools

There are excellent tools like Pylon [link] that consolidate support channels into a unified platform, perfect for B2B. They offer additional features such as tickets tracking, auto-nudging, auto-tagging, and more.  

Customer feedback analysis tools

There are also great tools that uses AI to analyze and aggregate customer feedback into accurate and actionable insights. They also go beyond analyzing tickets, but also conversations, surveys, and transcripts, helping you to understand customers better.

1. Thematic
2. SentiSum
3. Syncly
4. Monterey AI

Final Thoughts

There are some challenges we are actively addressing - e.g. how to effectively gather customer context in one go to reduce the back-and-forth, how to accurately infer product behavior changes from code changes in large organizations, etc. Tackling these hard challenges is what makes building exciting!

Top-notch B2B customer service builds trust and loyalty. Help your customers succeed, and you'll succeed too. It's that simple. If you'd like to chat anything on customer support, AI, or the best matcha spots in NYC, I'd love to connect! Feel free to schedule a chat [link] with me or DM me on LinkedIn [link]. 💛