[Un]Churned · 13 Jul 2026 · From the week of 13 July
The Old Playbook Is Obsolete. The New One Isn't Built. ft. Amanda Moran (Darktrace)
These are notes on the conversation, checked against its transcript. The episode itself has the full discussion.
In brief
Amanda Moran, VP of Digital Customer Success at Darktrace, speaks with host Josh Schachter of Gainsight about rebuilding customer success about six months into her role. She previously spent eight years leading digital CS at LinkedIn, where her team was part of a cross-functional initiative on personalized AI-driven email engagement. The episode covers Darktrace's digital-first and pooled support model for roughly 10,000 customers, renewals ownership, the data connection work she sees as a prerequisite for agents, and the in-between phase before autonomous agents are ready. Her central concern is building CS capabilities now without locking in manual automation that future agents would replace.
For founders
- Darktrace serves roughly 10,000 customers through a digital-first model for one segment, with a pooled team of CSMs, customer success engineers and support pulled in as needed, and about five people dedicated to the digital programs.
- Amanda Moran says connecting customer data across product usage, contract systems, communications, community and learning activity is the first step before agents can be used, and she calls it a manual first step.
- Amanda Moran is trying to avoid building CS workflows today that become manual automation she would later have to replace as agents mature.
- Darktrace's AI focus is moving to securing agentic AI starting around July, which she frames as a new layer of risk created by agents operating on their own.
For revenue leaders
- At Darktrace, CS owns the renewal targets and a renewals team comes in at the end of the lifecycle to get the contract signed and answer customer questions.
- At LinkedIn, personalized emails with next-best-action recommendations led to roughly 30% of engaged customers increasing use of one of three recommended features.
- At LinkedIn, the same email approach produced a 10% uplift in getting customers who had not logged in at all back into the product, customers she described as at increased risk of churn or seats at risk.
- Darktrace is standing up its first renewals outreach campaign now, and Amanda Moran hopes to have some renewals fully automated within the next year.
What was said 17, most useful first
At LinkedIn, AI-generated personalized emails with next-best-action recommendations lifted feature adoption for about 30% of engaged customers. Listen
The model used customer product adoption data, product best practices and how peers were adopting the product to generate personalized emails with next-best-action recommendations. The effort started around 2023 and was iterated on for efficiency and personalization. Roughly 30% of customers who engaged with the email saw an increase in one of the three features recommended to them.
“roughly 30 % saw an increase in one of the features, one of the three features that we were recommending to them.”
Listen to the episode Retention & customer success Link to this Report a problem
At LinkedIn, the personalized email approach lifted re-engagement among customers who had stopped logging in, with a 10% uplift. Listen
The team tested the same approach on customers who were not engaging with the product at all, to see whether they could be brought back. Amanda Moran reported a 10% uplift in those customers returning to the product. She said these were customers at increased risk of churn or on contracts where seats would have been at risk without use.
“we saw a 10 % uplift in customers who hadn't been coming in at all to getting them back in the product”
Listen to the episode Retention & customer success Link to this Report a problem
Customer community participation and learning activity could feed customer health scores as signals of where someone is in their journey. Listen
Amanda Moran asked what customers do in the community and at the learning center, and suggested those activities could be connected into a view of the customer. Josh Schachter said he had not considered using community participation, the flows customers go through, certifications and education speed bumps as health score signals. The idea is presented as an area to pursue rather than something Darktrace has already built.
“how do they use the community? What are they doing at our learning center?”
Listen to the episode Retention & customer success Link to this Report a problem
Darktrace runs a digital-first lifecycle for one customer segment and hands off to a pooled team on demand. Listen
For a segment of Darktrace customers, digital channels deliver education and product notices first. When a customer needs help, three teams can be pulled in: CSMs for value and usage issues, customer success engineering for technical issues, and support for break-fix issues. A sales team also handles commercial issues that come up mid-lifecycle for that segment.
“there will be a pooled model where we'll have basically three different teams that we pull in depending on what the customer need is.”
Listen to the episode Retention & customer success Link to this Report a problem
Darktrace has about five people dedicated to its digital programs, supporting roughly 10,000 customers. Listen
Amanda Moran described the digital program as foundational for all customers, with some parts targeted at particular segments. The community is meant to matter most for smaller customers who get less one-to-one engagement. She said the team is still being built and will expand as needed.
“there are about five people who are dedicated to driving the digital programs.”
Listen to the episode Retention & customer success Link to this Report a problem
Darktrace's digital CS team is organized into lifecycle program managers, each covering a phase of the customer lifecycle. Listen
Each program manager looks at the customer experience across every product a customer uses in that phase, since customers often have multiple Darktrace products. The aim is to orchestrate how customers receive information across the people and, eventually, the agents they interact with. A separate community program manager is building a customer community launching within a couple of months.
“We have a life cycle program managers. And so they are going to be focused on, and I'm building the team right now.”
Listen to the episode Retention & customer success Link to this Report a problem
Darktrace is standing up its first renewals outreach campaign and hopes to fully automate renewals where it can within a year. Listen
Amanda Moran said the team is just launching its first renewals outreach campaign to make that part of the process more seamless. She said the goal is to automate fully any renewals where that makes it easier for the customer and the renewals team, and she hopes some version of that will be in place within the next year.
“if there are renewals that we can fully automate to make it easier for the customer and the renewals team, that's what we're hoping to get within the next year”
Listen to the episode Retention & customer success Link to this Report a problem
Amanda Moran expects an agent that combines multiple risk factors to identify at-risk customers and send outreach at the right time and channel, though not yet. Listen
She described an agent that would connect customer data, identify risk from a combination of factors that people may not have the bandwidth to spot, develop outreach, and decide when and through which channel to deliver it. She said that version of digital outreach will arrive at some point but is not the version available right now.
“I think that version of digital outreach will be here at some point, but that's not the version right now.”
Amanda Moran wants to avoid building manual automation that she would later have to unwind. Listen
She is weighing what can be put in place now while building towards a future version of agents. Her concern is building in a way that does not produce a manual version of scale, so the team does not become locked into manual automation when it wants to build towards the agentic future.
“how do we think about building so that we're building something that is not going to be very sort of the manual version of scale”
Darktrace has created an AI Tiger team to coordinate how agents are developed across the business. Listen
Everyone at Darktrace is testing agents individually, with cross-team sharing about whether work can be done more efficiently or smartly. A new agentic Tiger team is launching to coordinate agent development, surface needs, connect agents and orchestrate work. Amanda Moran said that coordination will be critical as the company moves into its next phase of working internally and with customers.
“We also have an AI or an agentic Tiger team that is launching to coordinate the way that we're developing agents across the business.”
Darktrace is shifting its product focus to securing agentic AI, which it sees as a new layer of risk. Listen
Darktrace previously secured networks, email and cloud. Amanda Moran said companies are now using agentic AI and agents are operating on their own, which creates a new layer of risk. Darktrace's focus on securing AI itself starts around July.
“agentic AI as something that we are all using and that all businesses are now figuring out how we're going to use it is sort of the newest space that companies need help securing”
At Darktrace, customer success owns renewal targets, and a renewals team takes over at the end of the lifecycle. Listen
Once customers pass the middle phase of digital-first engagement and handoffs to pooled teams, the renewal manager comes in to make sure the contract is signed and to answer customer questions. Amanda Moran said CS owns the renewal targets.
“CS owns the renewal, basically. The renewal targets are owned by CS, and there's a renewals team that comes in at the end of the lifecycle.”
Listen to the episode Sales process & deals Link to this Report a problem
Connecting customer data across systems is the first manual step before agents can identify risk and act on it. Listen
Amanda Moran listed the data to connect: product usage, contract systems, all customer communications, community activity and learning center activity. She said LinkedIn had built a scaled digital program over many years but still had things that were not connected in the background. She described connecting this data as the manual first step toward using agents.
“the sort of manual, in a way, first step, I think is connecting all of those data components.”
Darktrace expects to connect data in phases, putting core components in place before adding community engagement to the customer view. Listen
Amanda Moran said data is one of the foundational problems the team is still working through. She expects to approach it in phases, first assembling core components and then adding pieces such as community engagement into what the team sees about a customer.
“we'll have to approach it in a like in a phased way in terms of getting core components together and then adding in these pieces”
Her team's near-term agent work is aimed at systems that help customers get value from the product. Listen
She said she is thinking about what agentic possibilities exist for digital CS in the near term and what the team wants a year from now. The near-term aim is a system that supports customers and helps them get value from the product, with agents playing a role where a version available today could make life easier for customers.
“How do we create the system that is going to support our customers, that is going to help us efficiently make sure that they are getting value?”
Josh Schachter expects agentic customer success tools to become fully autonomous, though the timing is uncertain. Listen
He said that at some point teams will be able to turn on agentic tools and have them do everything, but there is a current state where the tools are not there yet. He said teams do not want to spend time configuring workflows the way they used to, which points to an in-between model.
“two years from now, you know, whatever two or three years from now or five months Who knows the the agentic stuff will be fully autonomous”
Amanda Moran is asking how leaders should approach the in-between phase, after rules-based outreach and before autonomous agents. Listen
She said teams do not want to keep building the rules-based outreach they used to have, but are not yet at agents that can act as limited autonomous team members. She framed the open question as how to use AI to make outreach more efficient and personalized on the path toward agents that can be part of the team.
“what's the in -between step and how do we use AI to make the outreach more efficient, more personalized while we're on this path”