Layoffs have hit go-to-market teams hard, and most people respond the same way. They update the resume, open LinkedIn and start applying to everything that looks close. Even people who use AI for their job search usually just use it to write cover letters faster. It’s volume over strategy, which no marketer would accept in their own campaigns.
But according to Christine Castro, the fix is a different method, not more effort. The tools marketers use to find accounts, read buying signals and time their outreach work just as well when the product being marketed is you.
In this episode of Performance Delivered, host Steffen Horst is joined by Christine Castro, a B2B cybersecurity marketing executive with more than two decades of experience building and scaling marketing engines for pre-IPO security companies, from seed to Series C. At HYPR, her teams delivered 3x year-over-year ARR growth from marketing-sourced leads, and she has grown teams from 2 to 20 across the US, EMEA and APAC. Most recently at Constella Intelligence, she experienced a company-wide layoff, after which she built GTM Radar, an AI agent that tracks open go-to-market roles at pre-IPO cybersecurity companies. The two dig into how she built it and what any marketer can borrow from the method.
A job search is a pipeline problem, and you are the product
Christine’s layoff was purely financial, and she knew the usual reaction: jump into LinkedIn, message your network, apply. She chose to be more systematic and started where she would start any campaign, with an ideal customer profile.
“I took what I think about for a marketing campaign, and I started applying it to myself,” Christine says. Her ICP was pre-IPO cybersecurity companies that needed a VP of marketing or a CMO. Her buyers were founders and CEOs. “Instead of marketing a cybersecurity product, this time I’m the product.”
That reframe changes the work. You stop asking “what’s open?” and start asking which accounts fit, who the decision-makers are and when they’re likely to buy. In her words: “I just pointed the same stack at a different target.”
An AI job search agent starts as a spreadsheet
GTM Radar didn’t start as an agent. It started as a spreadsheet of companies Christine wanted to work with, then grew as she added research. The idea for automating it came from her last role, where she had directed an agentic AI SDR pilot using 11x. Her team was lean, had no SDRs, and needed to ramp up targeting, so she turned to AI to take over the tedious research.
She applied the same thinking to her own search, and the tool kept growing. At the time of recording, GTM Radar was tracking more than 614 pre-IPO cybersecurity companies and 3,660 open go-to-market roles across the US, Europe, the Middle East, Asia and Latin America, with new roles surfaced every night. Anyone can filter by function, seniority, location, remote status, and full-time or fractional work, then sign up for a daily email digest of the roles surfaced in the previous 24 hours.
Christine has kept it focused on cybersecurity on purpose. “I want to keep it just cybersecurity because that is the space that I know.” Narrow scope is part of what makes the signal useful.
Intent signals tell you who’s about to hire before the job is posted
Under the public job board sits a second layer that only Christine sees: a private CRM built for her own outreach. This is where the pipeline thinking does the real work.
The agent watches for intent signals: a hiring gap where a pre-IPO company has no marketing leader, a hiring burst, a recent funding raise, a recent job move or someone leaving the organization. For each target company, it pulls the executive team, the investors and the company’s specific cybersecurity category, then gives her a list of who to contact.
These are the same signals a good ABM program tracks. A company that just raised a round is about to spend. A company that just lost its marketing leader needs to backfill. You don’t need a custom agent to start watching for them. “Funding announcements in your target market are public announcements,” Christine says. “You can start there.”
Keep human oversight where the stakes are high
Christine could have taken the agent further. It could write her outreach emails, script her calls and send the messages automatically. She chose not to.
“I want to keep the human oversight part of this intact,” she says. If she were running a marketing program aimed at thousands of people, automation would make sense. Her target list is much smaller, and every conversation matters.
That’s a useful rule for any AI workflow: automate the research and the sorting, and keep a person in the loop where judgment and relationships decide the outcome.
The proof: a confidential role surfaced before the recruiter called
Building something clever is one thing, and getting results is another. Christine hasn’t landed a role yet, but she’s in several interview processes, and one story shows the system working.
A recruiter called her about a confidential search, a role that wasn’t posted anywhere. GTM Radar had already flagged the company through its intent signals, so Christine asked whether it was that particular company. The recruiter was shocked and assumed another recruiter had approached her. Christine walked her through GTM Radar. The company was actually next on her outreach list, and since a retained recruiter was already involved, she let the recruiter represent her.
The tool is also finding the audience it was built for. More than 100 people, including marketers, account executives and RevOps professionals, have signed up for the daily digest.
You don’t need an engineering degree to build an AI agent
Steffen notes that almost everyone uses LLMs in some way, but very few use them to build agents, partly because people assume it takes engineering skills. Christine isn’t an engineer. She built the whole thing in Claude Code on her own subscription, without a cybersecurity vendor’s budget, and ran out of credits every day while building it.
“I am not an engineer. I’m not a coder,” she says. She had built websites in WordPress and HubSpot, but could never have built the filters or an agent that searches on its own. What she had was the ability to conceptualize it. She could have handed the same directions to an engineer for thousands of dollars. Instead, she gave them to an agent on her laptop.
Her process was iterative. First Claude Code built the spreadsheet, about 150 companies and their contacts. Then she asked better questions: How can I find the right companies faster? How can I tell they’re hiring right now? The spreadsheet became a website, then an agent running a nightly cron job to find new companies and roles, then a system layered with intent signals.
The skill, she says, is “knowing what I need, being able to describe it very precisely, and treating AI as the builder rather than just a search box that you ask questions.”
Build for AI search, not just Google
Christine applied the same thinking to her own website. “SEO’s not quite dead yet,” she says, but people increasingly type questions into ChatGPT, Claude or Gemini and expect an answer back. “Sometimes they never even leave their ask engine. So if you’re not showing up in there, you’re already kind of out of the game.”
Every page on her site uses JSON-LD schema so machines understand what it’s about. The site has an llms.txt file telling AI crawlers what to read and how. And the content answers each question up front instead of burying it lower on the page. She took the same approach at her last company, rebuilding the website from scratch from the ask-engine perspective instead of reskinning it.
What to do Monday morning
For anyone who was just laid off, or is worried about what’s coming, Christine’s advice is direct: “Stop treating your job search as a job search and start treating it like a pipeline.”
Define your ICP tightly. “If you can’t say it in one sentence, you’re going to spray and pray.” Pick one signal to watch, because “you don’t need an agent immediately.” Then use AI to build something: start a spreadsheet, describe the tracking system you wish you had, and have AI help you build it.
And keep the strategy for yourself. “Don’t outsource your creativity,” Christine says. “Outsource the repetitive tasks or the things that you just can’t do.”
As Steffen puts it, when you treat a problem like a pipeline and use AI to build instead of just ask, you stop waiting for opportunities and start creating them.
Episode Transcript
Steffen (00:12)
Welcome back to Performance Delivered, Insider Secrets to Marketing Success, the podcast where we explore what's actually driving growth and performance in today's complex marketing landscape. I'm your host, Steffen Horst, and today we're taking the pipeline mindset somewhere most marketers never think to point it.
Steffen (00:29)
Their own career. Layoffs have hit go-to-market teams hard. And when it happens, most people do the same thing: update the resume, open LinkedIn, and start applying. It's volume over strategy. The exact approach we would never accept in our own marketing. My guest today did something different. After company-wide reduction in force, she took what she had learned directing an agentic AI SDR pilot and pointed it at the new target, her own next role. The result
Steffen (00:59)
Is GTM Radar, an AI agent that tracks every open go-to-market role at pre-IPO cybersecurity companies, watches for intent signals like funding rounds and leadership gaps, and delivers a daily digest. But this conversation isn't really about a job board, it's about a method. How any marketer, seller, or operator can treat a search like a pipeline problem and use AI to build instead of just ask. Joining me today is Christine Castro.
Steffen (01:28)
Christine is a B2B cybersecurity marketing executive with more than two decades of experience building and scaling marketing engines for pre-IPO security companies, from seed through Series C. She's led marketing at Hyper, Morphe's, Dusk Rice, Dog Control, Cognia, and most recently Constella Intelligence. Her teams have delivered three times year over year ARR growth from marketing source leads, and she's grown teams from two
Steffen (01:58)
to twenty across the US, India, and APAC. Today she builds in public at
Steffen (02:05)
Christine, welcome to Performance Delivered. It's great to have you on.
Christine Castro (02:10)
Thank you so much for having me. It's great to be here.
Steffen (02:12)
Now, Christine, let's start where this story begins. A layoff is one of the most disorienting things that can happen in a career. Most people go straight into reactive mode. You were running marketing at a cybersecurity company and got caught in a company wide layoff. Most people in that spot open LinkedIn and start applying. You built an agent. What happened?
Christine Castro (02:35)
The layoff was purely financial. It was it's happening a lot right now. And like you said, a lot of people do go into that sort of reactive mode where they just jump into LinkedIn and hit up their networks and start applying to things. I wanted to be a little bit more systemic in my approach and I started building out this spreadsheet of my ideal customer profile. Basically, I I took what I think about
Christine Castro (03:01)
For a marketing campaign, and I started applying it to myself. I know the type of company that I want to work in. I know how I like to work. So I started to build a spreadsheet that would inform me of potential openings and the who the hiring managers were at those specific companies and who I could reach out to. But I had also kind of conceptualized this and how I started to evolve.
Christine Castro (03:29)
what became GTM Radar because shortly before my layoff, I had directed an agentic AI SDR pilot. And basically what I did with that was I we we had a really lean team as it was, even prior to the layoffs. We didn't have SDRs. And I need to really ramp up what I was doing, how I was targeting. So I turned to AI, like a lot of us are doing right now. And I think AI is has such a great place in our work
Christine Castro (03:58)
flows. I love it to kind of get out from you know some of the tedious work, which is like the research and some of the the searching for who could be working within an organization and letting AI run as an agent. So I did the same thing with with my own job search. I built out what my ICP was which is pre IPO cybersecurity companies that needed a VP of marketing or a CMO.
Christine Castro (04:26)
And who my potential, you know, buyers are, and that would be founders and CEOs. And I am my own product, right? I am, you know, instead of marketing a cybersecurity product, this time I'm the product. So using the same tools that I would use as a marketer, I started to apply that to myself as a marketing leader. I just pointed the same stack at a different target.
Steffen (04:49)
So instead of applying,
Steffen (04:50)
you treated the search like a pipeline problem. Find the accounts, watch for
Christine Castro (04:53)
I absolutely did. Yep.
Steffen (04:55)
the signals and reach out when the time is right.
Steffen (04:57)
Now, which brings us to the tool itself. walk us through GTM Radar. What does it actually do? And where did it start?
Christine Castro (05:07)
So it started just as a spreadsheet. I was just tracking these are some companies that I would like to be working with. these are some in the space that I'm aware of. And then I started to do additional research. And then I really just decided why not take what I did with the agentic SDR pilot that I had run at Constella and apply it to myself. And I built out a full agent tool that now it's an actual job board. So instead of just tracking for myself, I I figured
Christine Castro (05:36)
this was something I could do is you know continue the evolution of the tool that I was using for myself. As of this morning, it's currently tracking more than six hundred and fourteen pre-IPO cybersecurity companies. And there's 3,660 open G go-to-market roles across the United States, all of Europe, Middle East, Asia, and Latin America. So it's really it's a global solution at this point and it's surfacing new roles.
Christine Castro (06:05)
every single night. it's also given me the ability to filter by function, by seniority, location, whether it's remote, if it's a full time or a fractional opportunity. And if others are using this tool, they can now sign up with a daily digest and set your own filters. Every morning you have an email delivered that has the roles that were surfaced the previous twenty-four hours.
Steffen (06:30)
So so is it is it still focused on cybersecurity or or is there an option to expand beyond that?
Christine Castro (06:36)
keep it to what I know the best. And cybersecurity in the pre-IPO space is what I know the best. I could add additional filters and add in different tech segments, different B2B SaaS segments, but I wanna keep it just cybersecurity because that is the space that I know.
Steffen (06:53)
Interesting. So so there's a layer running underneath that only you see. What is that layer and what signals is it watching for?
Christine Castro (07:02)
So yes, there's another layer that's underneath the public board. And that layer gives me a it builds out a CRM on my back end that I'm using for myself to do outreach. So, like I said, it's it's surfacing the roles that are posted, and that is what is publicly available on the GTM Radar. For the things that are at the intent layer, it is looking for things like hiring gaps if there is.
Christine Castro (07:30)
It's it's it will search you know pre-IPO company and if there is no marketing leader at that role, it will surface that company as having a hiring gap. If there's a hiring burst, if there is a recent funding raise, if there is a recent job move, if somebody leaves the organization. It treats those all as intent signals and then it provides a list for me for who I can reach out to.
Christine Castro (07:57)
I could go a further step and have it write me emails and have it write me a script to do the outreach. And I could also automate that outreach. I chose not to because I want to keep the human oversight part of this intact. but it I could do that. And certainly that would be something that could be looked at if I wanted to apply at scale or if I was applying this to a marketing initiative where I needed to be able to.
Christine Castro (08:23)
do outreach to thousands of people.
Steffen (08:26)
Yeah.
Christine Castro (08:27)
my targets are a lot smaller. So I want to keep myself in that the human oversight role here. So that's where I'm I am with it. But that that layer is not a layer that I share. So it's building out a CRM just for me, just where I can start to approach these people. It raises the entire executive team, the investors for the company, what you know, the space that they're in specifically.
Christine Castro (08:53)
The the category of cybersecurity, etcetera.
Steffen (08:56)
Interesting. Now building something clever is is one thing, right? Getting results from it is is another. So far, does the system does the does the AI agent work? what have you gotten so far out of it?
Christine Castro (09:13)
It absolutely does. I haven't landed a role yet, but I'm in the process with several interview processes and I'm hopeful that it will quickly I'll quickly land. I think what it's it's definitely working. I'm surfacing the right roles. a funny anecdote is that it surfaced one role and then this it was not posted anywhere. It is a a confidential search and
Christine Castro (09:39)
I was able to figure out when the recru a recruiter called me about it. Like I said, not posted anywhere, but my tool had surfaced it with the intent signals. And because of that, I knew exactly who it was when the recruiter reached out to me. So I was able to say, Hey, you know, just out of curiosity, is it's this this particular company? And she was shocked. She's like, Did you get approached by another recruiter? Like I think she thought that, you know, perhaps multiple recruiters were working on this and I said no. And and I
Christine Castro (10:08)
walked her through the GTM Radar and she was like, that's really impressive. Like you knew exactly who this company was just based on a couple of bullet points in the job description. And it was because the my tool had uncovered it and sourced it. And they were actually next on my list of companies to reach out to. I did
Steffen (10:29)
interesting.
Christine Castro (10:29)
just have the the recruiter reach out on my behalf because you know you get in a lot quicker and if they already have a retained recruiter.
Christine Castro (10:38)
certainly makes
Steffen (10:38)
Yeah.
Christine Castro (10:38)
more sense to go that way. But it's the tool is absolutely doing what it's intended to do. And there's
Steffen (10:44)
That's great.
Christine Castro (10:44)
also, I think, you know, the the one other data point that I can point to is there's over a hundred people that are now signed up for it on the daily digest. So there's hundred people that are other marketers, account executives, RevOps folks, they're also using this tool. So I think it's definitely getting the audience that
Christine Castro (11:05)
It was intended for, but it's also working for me in my search.
Steffen (11:09)
Yeah. So as you said, there's a layer that only you have access to. What do people that that sign up for the digest that that use the tool what do they get out of it? Where where does it stop?
Christine Castro (11:22)
Yeah, so they can select their filters. you know, for example, my filter would be marketing at the C suite or the VP level and remote roles. But you can go further down to you can do specific marketing disciplines. If you're looking on the sales side, you can do account executives, regional directors at all different seniority levels. You really select your own filters and then when you submit your your digest sign up, you'll get an email back with
Christine Castro (11:49)
Every new role that was surfaced in the last twenty four hours. So once a day you'll get the the email sent directly to your email with every new role that has come up in the previous twenty four hours.
Steffen (12:01)
That's great. Now, I feel like everyone these days uses AI LLMs in in some shape
Christine Castro (12:10)
Me?
Steffen (12:12)
or form, right? I think a very small amount of people are actually using it to build agents, to to kind of take care of specific parts of their day-to-day. And I think a lot of people think that you need actually an engineering.
Steffen (12:26)
Degree or you need to understand how to engineer something to build an AI. but you're a marketer and not an engineer. How did you how did you actually build this? I mean you talked about that you you built an SDR solution for your previous company, but but how did you approach this specifically to kind of turn your Excel sheet? This is where we where you started, as you explained earlier.
Steffen (12:52)
into an agent that now does a lot of the work and takes off a lot of the the the the kind of the manual part that that usually has to be done by going into LinkedIn, Indeed and all the other job boards and compiling the information.
Christine Castro (13:08)
So the the SDR pilot that I ran when I was still at Constella, that was with an actual tool, 11x. Fantastic tool, and it it certainly gave me the idea for how I started to build my own tool. but my own tool was built completely in Claude Code. I do not have kind of a the a budget of a cybersecurity vendor, so I had to be really creative in what I was doing. So
Christine Castro (13:35)
I do have a Claude Code subscription and that was what I used. And yes, I did run out of credits every single day as I was building this. but the whole thing, I think one of the things that makes utilizing tools like this great is that it's a constant evolution. Every time I had a new idea for additional features and things that I could add, I could put in the prompts and and start to get additional filters, additional links, additional tools that
Christine Castro (14:04)
really help me with this. And I think that is what is really great about leveraging AI for things that I couldn't have done. I am not an engineer. I'm not a coder. I've built websites in the past and things like WordPress and and HubSpot, but I could never have built out some of these filters. I could never have built out an agent that goes searching for something. That's just not my my wheelhouse. but I know how to
Christine Castro (14:34)
conceptualized. Like I could have given these directives to an actual engineer. It would have cost me thousands of dollars. but I I could have done that. I think so I think where I the big benefit is is that I was running tools like this and I've been doing you know programs like this for a long time within my career. But the difference now is I have my own agent sitting in on my laptop that I can give the prompts to.
Christine Castro (15:04)
and and really to be able to start to grow from there.
Steffen (15:08)
Okay.
Christine Castro (15:08)
I think the skill set is is knowing what I need, being able to describe it very precisely, and treating AI as the builder rather than just a a search box that you ask questions. I don't want
Steffen (15:20)
Yeah. Yeah.
Christine Castro (15:21)
to leverage it just to ask questions, which you know it certainly has a purpose just for the questions, but I wanted it something that I could really use as a builder.
Steffen (15:30)
That makes that makes sense. Now if if you if you allow me, I'd still like to kind of go back to, you know, you have you have your Excel document. What is the next step? How how where do you start? Because I think a lot of people it's like, I don't even know where to start. Yeah, I have something jotted down at a piece of paper, but what's the next step? What am I going to do once I go to Claude Code in order to get to some some product that can do something?
Christine Castro (15:58)
So I basically asked Claude Code to start to build me a tool rather
Steffen (16:04)
Mm-hmm.
Christine Castro (16:04)
than just a spreadsheet. So initially it built out the spreadsheet. I think there was about a hundred and fifty companies on there, who the
