Decimal Raises $4M to Bring Engineering-Level Understanding to Customer Support

Today Kevin Cherian and I are beyond excited to announce our $4M seed round for Decimal, co-led by Khosla Ventures and Kearny Jackson to bring engineering-level understanding to every customer interaction.

Since inception, we’ve partnered with some of the best AI-native companies: Granola, Resilinc, Omnea, Tealium, BuildOps, and Lucidworks to help their support and engineering teams resolve customer problems. Working with these teams has cemented our belief that the most valuable issues to solve are also the most difficult, complex, and technical.

When a customer reports an issue, they do it based on the symptoms they see. Let’s say some data that a customer expects is no longer getting populated in their dashboard. That’s probably all they will say. But the support agent needs to answer why. Is it a temporary blip? An incident the customer wasn’t aware of? An error with a configuration?

From the support agent’s point of view, this report could be a simple misunderstanding, a bug requiring engineering work, or anything in between. To properly diagnose a ticket, a support agent can rely on help center documentation or playbooks for the most common issues, but anything deeper requires gathering and synthesizing information from several sources: logs, code, configs, a reproduction of the issue, and so on. At this point, the agent needs to conduct a full-blown investigation.

All of this takes up a ton of time, and missing even one minor piece of context can lead to an avoidable escalation to engineering. Even worse, the engineer often has to do the investigation all over again, taking up valuable time from multiple teams.

At Decimal, we believe that these are the most important issues to address. They require combining information from many disconnected systems, and which aren’t solvable by just linking an article or sending a canned response. The first wave of AI support products focused on finding answers faster: search the knowledge base, retrieve similar tickets, and quickly generate a response. This is useful for questions where the answer already exists. However, to tackle the hardest customer problems, someone has to figure out the answer, not just find it.

That’s why we’re building Decimal! Whenever a ticket is first seen, Decimal can deeply investigate the customer’s issue, taking into account everything it has access to. Then, it can determine the most appropriate action(s). For cut-and-dry problems, Decimal can just reply directly to the customer. Otherwise, it can provide the human handling the issue with enough context to greatly reduce their workload.

With the aid of AI agents like Decimal, a new category is emerging: Customer Engineering. A customer engineer owns the customer interaction end-to-end. They are on the ground communicating with customers, while also doing deep technical dives into the product. Decimal helps with the resolution end-to-end, whether it’s updating docs, opening a bug fix PR, or just formulating the perfect response.

With a lot of the work done before a human ever touches a ticket, customer engineers can focus their energy on the truly complex issues, or just adding a human touch to a response. Rather than spending a bulk of their time gathering and synthesizing info, they can focus on the places where they can really drive an outcome.

So what does the world look like in an org following the customer engineering model? The customer engineer owns the entire customer request lifecycle. They can resolve all sorts of issues, including ones requiring deep technical investigation. The customer engineer knows the customer best, and can add valuable human insight to all interactions they touch.

The engineers, on the other hand, can focus all of their time on building, rather than reproducing customer issues or investigating the root cause of a customer problem. This would all be done by the time it reaches engineering, making their jobs easier. Additionally, only tickets requiring extensive engineering work will actually reach their table, reducing the amount of interruptions and context switches.

We believe the future isn't one where AI simply answers more support tickets. It's one where it enables the teams closest to customers to understand what happened, investigate the root cause, take action, and ultimately own the resolution.

If this sounds like an interesting problem to you, we’re hiring!