Meet the Agent: How Our Deployment Strategy Team Built Clara to Scale With Our Customers
It's 9:58am. My next customer call starts in two minutes.
Instead of scrambling through old notes, checking messages, searching through customer success management (CSM) tools, and trying to remember the last conversation I had with this customer, I open a single document — created by Clara.
Clara creates briefs to help me confidently walk into conversations with customers, even if it’s a wall-to-wall meeting day. These briefs include the customer’s account history, their recent news, names and important details about the folks I’ll be speaking with, open product questions, any risks I should be aware of, and even talking points tailored to our meeting. So by the time I join the call, I'm ready.
Clara also goes to work after these customer calls end. Within minutes, she drafts a concise conversation summary, generates and assigns action items (organized into what TRM owns versus what the customer owns), prepares a follow-up email, writes internal messages for cross-functional coordination, documents the customer’s feature requests, and updates our customer success tools.
By taking care of the busywork, Clara gives me more time to focus on what matters most: advocating for my customers.
The hidden administrative burden
As a deployment strategist, my days are filled with customer meetings, internal discussions, and cross-functional collaboration. Some days my calendar is booked solid from morning to evening, and it always feels like there's another meeting just around the corner. The meetings themselves aren't the challenge; it's everything that happens before and after them.
My account portfolio spans DeFi startups to traditional financial institutions, each with different priorities, histories, and stakeholders. Preparing for each conversation means digging through notes, reviewing previous meetings, researching industry developments, coordinating internally, and making sure I understand what matters most to each customer before joining meetings.
Then there’s the follow-up work, which is just as important as the preparation. Rich insights are often found in one call, and it’s my job to ensure they get across TRM in the fastest way possible. Every customer conversation creates action: documenting notes, updating customer success tools, coordinating with our product and engineering teams, sharing customer feedback, following up with teammates, and making sure nothing that was discussed live falls through the cracks. While these tasks aren’t particularly difficult, the time it takes to handle them all can add up quickly.
Making the most of customers’ precious time
Customer conversations are the highest-value part of my job. Every minute I spend chasing notes or writing follow-up emails is a minute I'm not spending helping a customer solve a problem. That's why I built Clara.
Clara was designed to take on the administrative-heavy work I was spending hours each week doing so that I could be freed to spend more time in those high-value customer conversations. Today, she saves me 30–45 minutes on every customer meeting. That adds up to hours every week that I now get back to spend with customers instead of doing admin tasks.
I like to think of Clara as another member of our team. She exists to help create more room for the human side of customer success, and enables me to focus on being present with my customers.
While Clara is my my assistant, co-pilot, and compass — drafting briefs, follow-ups, and other important work — I still stay in control. Every recommendation, email, feature request, and internal update still passes through and must be approved by me before it's shared.
How Clara works
Clara is an AI skill that anyone at TRM Labs can use. Her guiding motto is simple: Finding action in any direction.
Under the hood, she's built on a network of MCPs connected to as many of TRM's live systems as possible. The more context she has, the more helpful she becomes.
Today, Clara runs when I ask for her. Before a meeting, I'll simply say, "Prep me for this call" or "I have a customer meeting coming up." She acts as a research assistant, collecting information from multiple systems and presenting it in a way that's easy to consume. By gathering everything I need into a single story, Clara helps me walk into the conversation informed.
After the meeting, I'll ask for a post-call action plan or have her pull together my notes. Clara organizes every action item into what I own versus what someone else owns, drafts emails, prepares internal updates, and suggests feature requests based on customer feedback received during the call. She packages it all up into deliverables that I review before anything gets shared.
Clara also includes a built-in feedback loop where teammates can rate her responses and suggest improvements. One of the most rewarding parts of building her has been watching other people shape how she evolves, with every piece of feedback making her a little more useful for the next person.
What we’re learning and tweaking
As helpful as Clara is, she can sometimes be a little trigger-happy.
Sometimes Clara finds action where I don't think there is quite as much — for example, suggesting an action that doesn’t really warrant our focus or attention. And that’s okay! We're still learning together; and that’s why keeping a human in the loop is so important. Our goal is better outputs, not more outputs.
I’ve also discovered that one of my favorite use cases for Clara is preparing for meetings with multiple internal stakeholders. Instead of creating separate briefs for product, compliance advisory, customer success, and everyone else joining the conversation, Clara helps me build one document with dedicated sections for each team. I simply share the link, and everyone walks into the meeting with the context that's most relevant to them.
But perhaps the biggest surprise of all in the process of creating Clara was how much I — a non-technical person — could build on my own. I'd been doing all of this work manually for months because I assumed building an AI skill required engineering experience that I simply didn't have. So even though I was using AI every day, I still had to manually provide Claude context every single time I needed to prep for a customer call.
In process of turning Clara into a scalable skill, I went from asking, "What is an AI skill?" to shipping something that teams across GTM can now use, thanks in large part to the support of our incredible AI enablement team.
That shift completely changed how I think about AI. I stopped seeing it as another tool I had to use and started treating it like a teammate I could build.
From call to action, to taking action
Today, Clara waits for me to ask for help. Tomorrow, I want her to be proactive. My goal is for her to know when I have a customer meeting on my calendar, prepare my briefing automatically beforehand, recognize when the meeting has ended, and have my follow-up work ready before I've even switched tabs.
This next chapter — automation, personalization, and continuous learning — will still keep the core brief in mind: freeing me up to focus fully on my customers.
Because at the end of the day, customers won't remember how quickly I sent them a follow-up email. They will remember how well we listened, how well we partnered with them, and whether we helped solve their problems. If Clara gives me more time to do that, then she's fulfilling her mission.
{{horizontal-line}}
This post is part of a series where we introduce the AI agents TRMers have built. Each one has a name, a role, and a story. If building AI systems that make more room for time with customers sounds like your kind of work, we're hiring.
