This article may contain affiliate links. If you sign up through one, the author may earn a commission at no extra cost to you.
These are the four tools we use at Frontal, what each one does for us, and how they fit together in a go to market workflow.
Clay
The first tool is Clay. If you’ve ever built a prospect list manually, you know how messy this can get. You’ve got LinkedIn open, a company website, a database, a spreadsheet, and whatever tool you’re using to send your outreach. You find one piece of information here, another one over there, and then you spend most of your time moving between all the different tabs. At Frontal, we treat Clay as a data orchestration tool. What that means is that we use it to bring different sources of information together and then decide what to do with all that data.
Waterfall enrichment
One use case with Clay is waterfall enrichment. Say you want to find an email address and the first provider can’t find the one you need. You can have Clay check another provider, based on the workflow you’ve configured. This improves your chances of finding the information. Of course, it doesn’t mean that every record will be complete or that every email you find is ready to contact. But where Clay gets really useful for us is when we go beyond finding someone’s email and start researching the account itself. In our experience you can find a very wide range of data points with it.
Worth reaching?
Here you can see a workflow called wedge agent feeder. It gathers the data points we need to decide if a company is worth reaching out to. In our case we want to know if they’re running ads on LinkedIn and if they’re running Meta ads. We want to know their SDR headcount. And we want to check their email infrastructure and see if they already use a different domain to send outbound. We also run an ICP and time estimate. With all of this, we can decide if the company is worth contacting.
For us, a company is usually worth reaching if they’re not running LinkedIn ads, they’re not running Meta ads, but they have a high SDR count and a really bad email infrastructure. That’s prime time for us to reach out, because we can probably help them with a go to market strategy.
Railway
The second tool is Railway. Railway is a hosting platform. I’m not sure we can call it a tool per se, but I’m including it because once you have built something with AI, you need a place to deploy it so other people can use it and see it. Here’s an example from our business. Kenny, my business partner, built a generator that researches a company and uses Claude to produce a personalized GTM playbook. The idea is to give someone in your ideal customer profile something relevant to their business before asking them to book a call. On our website, we have a form where you enter your email address and your company website to request that playbook.
Here is one we generated earlier for Frontal itself. If you go to the blog, type in your email and give your website, we generate a playbook that breaks down all the outbound plays the company could run. For us it’s a nice way to get a foot in the door and show expertise, and from the lead magnet people can book a 30-minute meeting. Railway runs the application behind that experience. Here is the project, and you don’t need to understand all the boxes. There’s a small database and some agents running in the background.
We’ve taken a piece of expertise and turned it into something a prospect can request and use. You could apply the same idea to a calculator, an audit, or a research report for your own market. Start with a question your buyers care about, then build something that helps them answer it. An email capture form is only useful if the thing behind it is worth requesting.
Book a call
I know that learning all of these tools can feel overwhelming. If you’re running a business, you probably don’t have time to figure out every connection and every tool by yourself, and that’s where we can help. At Frontal, we help B2B companies build and run the go to market system across go to market engineering, paid media, and content. We can look at your current setup and where it makes sense to improve it.
Claude + context
The third tool is Claude. You can already use Claude to research a topic, review a document, or write a first draft. The part I want to show you is how useful it becomes when you give it the context and instructions for the way your business works. Kenny has been building a GTM operating system for our team. It brings together all the working context about the clients, the tools and the workflows, plus reusable instructions and the tools to do the work. Here you can see the Frontal go to market OS. We’ve been building it over the last few months, and every new go to market engineer who starts at Frontal now gets equipped with it, so they can do their work much better.
Quick demo
You can run it inside Claude Code or inside Codex. For this demo, I’m going to show you what it looks like. I’m using Kenny’s GTM OS, which we’ve already imported from GitHub, and running a quick demo with maybe 10 prospects to show how we would run a campaign for Frontal using the GTM OS we use internally.
You can see all the different steps here. We’ve used different skills to pre-train our Frontal OS. First it figures out the clan file and the current positioning, with Frontal as the example. It works out our ICP and our positioning. Based on that, it finds the top 10 accounts worth reaching out to, using a signal engine to pick the 10 we should focus on. Then it picks the right person at each company, chooses the angle for the copy, and gets the verified email. After that, it writes the entire sequence.
Email we would send
For each account, you get the company, the position we should target, and why now: the event, trigger or signal we can use. Here is the email we would send. There’s a new BDR starting, so we’d reach out to the CRO, Paul. “A first BDR tends to lose month one to domain warming up and a list nobody trusts yet. Loves market or signal engine and pull the 50 accounting buying signals this month, each with the reason. Want me to send it over so they can start on day one?” We use the fact that they’re hiring a BDR soon, we show expertise, and we give free value before asking for a meeting.
This is the idea of building reusable knowledge. Kenny, my business partner, is in my opinion one of the best people in the world at GTM, and we’ve compacted much of that knowledge directly into the GTM OS. Most of our go to market engineers don’t need to rethink everything whenever they start working for a new client.
Wispr Flow
The fourth tool is Wispr Flow. This one is much simpler. It lets you dictate into the place where you would normally type, and it lets you go much faster. When I’m giving AI instructions, the useful context is often already in my head. I need to explain what we are trying to achieve, who we are targeting, and what I don’t want it to assume, and writing all of that out is really time-consuming. For example, I can say: review the list’s summary for the B2B campaign. Tell me which records still need verification, what information is missing, and what we should check before using the contacts on LinkedIn.
Style
A cool thing about Wispr Flow is that you can choose the style it applies to different messages. You can have one style for your private messages and another for your business messages, and you can pick formal, casual or very casual. Sometimes when you dictate, you’re going to mumble and it won’t transcribe very well. If you double press, it starts recording, and then I can just take my hand off and chat almost as if I were dropping a voice note to a friend on WhatsApp.
Transform feature
That raw transcript would be very annoying to read. With the transform feature, all the filler words are removed and you get a succinct message. For me, it saves so much time. You also get insights: how fast you talk per minute, all the days you’ve used Wispr Flow, which apps you use it in the most, and how many fixes it made.
For me it’s one of those apps I could never go back from. Wispr Flow is one of my absolute favorites, and I think I’m much more productive with it. On our website, frontal deso, I literally used Wispr Flow to build the entire thing. It was just a GitHub repo, and I didn’t type a single line of code. Everything went through Wispr Flow to Claude Code to end up with a functioning website.
All four together
How would I bring those four tools together? Let’s say we want to start more relevant conversations with a specific type of B2B company. First, I would define the audience and the problem we can help them solve. I can use Wispr Flow to get all of that context into a written brief, then work through Claude Code to make it real. Next, I would use Clay to bring together the account research we need, including advertising activity, team information, and other signals relevant to our offer.
Then I can have Claude review that research and draft a useful angle for the conversation and the messaging, using the instructions and the evidence we’ve collected about what campaigns and copy work. If sending a personalized resource makes sense, I could build an application hosted on Railway and turn that idea into something the prospect would use and care about, exactly like the playbook I just showed you. Those are the roles each tool can play. Connecting the whole process still takes setup, and you need to decide where a person should review the output.
Why 25 first?
Before scaling it, I would always test on a small scale before automating anything. This is something I tell our team and clients all the time. Why would you start an outbound campaign to 5,000 people if you haven’t made it work manually for 25? Use those first manual conversations to find out whether the problem matters, whether your offer makes sense, and whether the resource is useful. When it’s working at a small scale, then you have something worth automating.
Clay coordinates the data. Claude does the research, build and review. Railway hosts the application. Wispr Flow helps us move faster. Start with the workflow your business needs and choose the tools that help you make it work.
Needs Review
- “clan file” in the Quick demo section: likely a transcription error (possibly “plan file” or “CLAUDE.md file”), but no confident match, so the original was left.
- “frontal deso” in the Wispr Flow section: probably the site’s domain (e.g. “frontal.so”), but it can’t be confirmed, so it was left as written.
- The sample email in “Email we would send” has garbled text: “Loves market or signal engine and pull the 50 accounting buying signals this month.” It likely means something like “Let us run our signal engine and pull the 50 accounts with buying signals this month,” but this was not changed. “BDRC” was corrected to “BDR” and quotation marks were added.
- “wedge agent feeder” (Clay workflow name) and “ICP and time estimate” were left as spoken. They may be misheard.
- “Kenya” appeared once for the business partner and was changed to “Kenny” to match the other mentions.
- The disclosure at the top is generic. Confirm whether any of the tools are actually affiliate links and adjust the wording if needed.
- The final Wispr Flow claim (building the entire website without typing code) is the speaker’s own account and was left as stated.
