GTM Engineering: The AI Revolution in B2B Sales

Meet Clay - The mastermind behind the AI Revolution

Clay is a GTM (Go-To-Market), sales and marketing platform founded in 2017 that had an astonishing growth in revenue and valuation. The company achieved a tenfold increase in revenue during 2022 and 2023, followed by a sixfold growth in 2024. After securing its first Series A funding in 2019, Clay reached a valuation of USD $1.25 billion five years later, during its most recent Series B funding round in 2024. According to Clay’s website, the platform is already utilized by over 5,000 customers, including prominent companies such as OpenAI, Canva, Anthropic, Ramp, and Rippling.

Yes, you heard it right. The AI founders themselves – OpenAI – uses Clay. As mentioned in this Forbes Article:

‍

The Head of GTM systems in OpenAI, Keith Jones, said that his team used Clay in its workflows to launch and test new business-to-business sales tactics on a daily basis. “In my professional opinion they have one of the most practical and exciting applications of AI, in a decades-old practice that has long been stale,” Jones wrote in an email.

‍

Clay is disrupting the way sales operates and it has potential to top up with the biggest sales and marketing platforms, like Hubspot and Salesforce, in a really (and let’s emphasize this “really”) short time.

Quoting the Forbes Article once again:

‍

Already handling the licensing and collection of outside, third-party data for customers, Clay has capitalized on improvements in AI to supercharge those efforts. Clay’s software can now do some of the work for the user, predicting what data points and patterns might be valuable. Its AI agent, called a ‘Claygent’, can take a potentially valuable but complicated question, such as finding every previously disclosed customer of a company, or every person on LinkedIn who worked in certain roles at that company over a specific time period, and return just the results.

‍

Let’s dive deeper.

‍

Clay revolutionizes the automation of processes that were previously impossible to automate. Before Clay, SDRs and sales professionals conducting cold outreach faced two primary approaches:

‍

• Sending bulk automated cold emails with highly templated messages, such as, “Hi {first name}, I wanted to reach out to {company_name} because we can help you achieve X result.” Unsurprisingly, 98% of these emails are ignored, as they are clearly generic and lack any real relevance to the recipient.

‍

• Employing a large team of SDRs and BDRs to manually research companies and craft personalized one-to-one emails. While this approach is more effective, it is costly, time-intensive, and heavily reliant on the size of the team—making it challenging for smaller teams to scale their outbound efforts.

‍

Now, everything has changed. Clay enables sending one-to-one personalized messaging at the scale of high-volume automated emails. This shift is transformative. Instead of needing a team of five SDRs, a company can now deploy one Go-To-Market (GTM) Engineer (more on this role later) who, with Clay’s setup, can send up to 30,000 emails per month. These emails are not only high in volume but also hyper-personalized, heavily researched, and highly relevant - appearing as though they were crafted by a person.

However, it’s important to note that fully AI-generated messages are not yet entirely viable for outreach. With Clay, users can create a hybrid approach: combining human input with AI-generated snippets to enhance templated content.

Beyond email personalization, Clay eliminates the tedious research process traditionally required of SDRs and BDRs. Previously, sales professionals would individually research prospects—scouring Google, analyzing 10-K filings, and leveraging LinkedIn Sales Navigator to gather data. This manual process was time-consuming and labor-intensive. Clay automates this entire research component by using AI agents to source publicly available information about individuals or companies, at scale. This allows sales teams to focus their time on selling rather than researching and writing.

Clay also excels in managing inbound lead generation. It centralizes all inbound leads from various marketing initiatives—such as LinkedIn ads, organic LinkedIn content, newsletters, lead magnet campaigns, webinars, and in-person events—into a single hub for scoring, enrichment, and normalization.

If your marketing efforts generate leads through multiple channels but you fail to act on them on a timely manner, you are wasting both time and resources. Clay ensures that all leads from these campaigns are automatically consolidated in one place, enabling you to act on them - fast and at scale. Tasks like enriching leads, gathering additional data, scoring, and qualifying can be seamlessly automated for both outbound and inbound efforts. Once organized, these leads can be transferred to platforms like Apollo, HubSpot, or Salesforce for outreach sequences.

The true power of Clay lies in its role as a connective platform that integrates with other go-to-market tools rather than competing with them. This eliminates the need for additional subscriptions while enhancing cost-efficiency.

‍

‍

Clay represents an unprecedented evolution in the application of AI within sales frameworks. If this innovation excites you as much as it did when I first encountered it, just wait a bit. There is much, much more. Let's take a deeper dive into AI-powered lead generation and automation possibilities - not just with Clay but also with AI GTM Agents - along with real-world case studies showcasing these tools in action.

‍

Precision Targeting

‍

Buying Intent Signals

Signals provide insights into a prospect’s “buying intent”. In other words, they help determine not only whether a prospect is a suitable match for the solution you are offering but also their “sales readiness”—whether they are likely prepared to engage with your sales team and if the timing is right for you to make a move. Examples of effective buying intent signals include:

‍

• Interactions on your website: content conversions, number of pages viewed, specific pages viewed (ex: pricing or contact page), forms submitted, etc

‍

• Email engagement: emails open, emails click, emails reply, and negative metrics such as “unsubscribes”.

‍

• Social media engagement: following your company on various channels, engaging with social content (shares, comments, etc)

‍

The examples above represent signals derived from traditional marketing analytics and tracking codes—commonly referred to as “first-party intent data.” This data is collected directly from your own digital environments, such as your website, CRM, social media efforts, or offline interactions. Now, we can explore a broader range of signal possibilities through “third-party intent data.” This data is gathered from external sources, offering a wider perspective on buyer intent by leveraging methods such as cooperative data from publishers or web scraping. Essentially, any publicly available information on the internet can be transformed into a buying intent signal. Examples of such signals include website content updates, financial filings, Google Maps reviews, or social media activity.

One commonly used example is “hiring intent.” If a company has open job postings in a specific area, it may indicate a need for related services. For instance, a company hiring a digital marketing analyst might require marketing services, while hiring for the HR department could signal the need for recruiting agency support. More complex examples are also possible: if you offer a compliance platform, you might monitor when prospects announce plans to pursue SOC 2 certification or open offices in highly regulated regions. Successful teams often combine multiple signals - for example, layering third-party signals (e.g., a company doubling its engineering team) with first-party signals (e.g., a trial user signing up). The possibilities are limited only by your creativity.

The image below illustrates the vast range of buying intent signals available within Clay. Additionally, Clay integrates with other data providers to further enhance its capabilities. For instance, the platform Delivr offers buying intent insights at an individual level rather than just at the company level. Common Room has compiled an impressive list of 100 signal ideas in a Google Sheets document - in case you need some inspiration.

‍

‍

Account and Persona Prospection

We are accustomed to filtering both companies and personas when building a list of prospects for outbound efforts, typically using criteria such as industry and company size for account-level information, and job titles for persona-level information. With Clay, however, we now have access to an expanded array of filters for both individuals and companies, along with an entirely new category for precise targeting, using "signals".

With the ability to access and retrieve virtually any publicly available data point on the internet, we can now leverage our creativity to develop new filters when searching for a company or an individual. For instance, when searching for a person, you can not only filter by job titles but also by their years of experience or specific keywords in their profile. If you have an Industrial Technology (IndTech) startup, you might search for terms like “Industry 4.0” or “Innovation” within a prospect's LinkedIn profile. Additionally, you could use an AI-powered Boolean logic formula to determine whether the individual is genuinely a decision-maker within their organization.

When searching for a company, you could create a prompt for an AI assistant to identify whether a business operates as B2B or B2C (a feature already available on platforms like Apollo) or even narrow it down further to determine if they are a SaaS company. For example, you could use a prompt such as:  

‍

“Is the company with the name {{Name}} and domain {{Domain}} a Software-as-a-Service company? Return two fields: the first as a boolean (\true\ for yes or \false\ for no), and the second field with the reasoning behind the response"

‍

The real deal, though, is the new ability we have now to search for small businesses using Google Maps. These are companies that are not on LinkedIn and up until now to scrape the web for this information was a painstaking task. Now we can prospect bakeries and hair salons just as easily as prospecting fortune 500 companies.

Follows some other examples of searches we can do now with Clay:

‍

‍

Data Enrichment

Until recently, I relied on two separate tools, Apollo and Lusha, to retrieve prospect phone numbers. I noticed that Lusha often provided alternative numbers for the same prospects, increasing the likelihood of reaching someone during cold calls. However, this process was highly time-consuming, as it required manually navigating each LinkedIn profile. Fortunately, those days are over. Clay offers a streamlined solution by allowing users to search for work emails and phone numbers across multiple data providers simultaneously, including Prospeo, DropContact, Datagma, Hunter, PeopleDataLabs, Nimbler, Apollo, Lusha, Snov, and others.

Well, maybe not simultaneously. Clay operates using a waterfall system, sequentially querying each source and stopping once a valid email is found. Users can decide whether to stop the search after finding an email - ideal when email addresses are the primary goal - or to continue enriching the data table with all available options, such as multiple phone numbers.

‍

‍

As illustrated in the image above, Clay’s email waterfall not only identifies emails but also validates them. Additionally, the waterfall system can enrich other data points, such as company-related information (industry, company size, revenue, funding stage) or other valuable insights like job openings, recent news, and additional data that provide a more comprehensive understanding of our prospects.

One of the most notable innovations in data enrichment is Clay’s ability to utilize its “AI Clay Agent.” If a specific data provider app is unavailable for certain information, users can create a custom prompt for the AI to search for the required data across the web. Anything, anywhere. In the example below, we are instructing AI to locate case studies on a prospect’s website. This information can then be leveraged to craft personalized emails, which is the topic we will see next.

‍

‍

Personalization at scale

The idea of this article is not to be a Clay tutorial. So let’s keep it simple just so you can understand the logic behind how this personalization at scale works. Once we have enriched our prospect’s data with the necessary details for personalization, we proceed to create “AI snippets” for each sentence in our email. Rather than generating full paragraphs or extensive content (as AI is not yet fully reliable for that), we carefully craft individual sentences or smaller segments of text, one at a time.

When developing these snippets, we provide clear, step-by-step instructions to guide the AI on how to construct a sentence using the relevant data from your enriched table. Example:

‍

“Always begin the subject line with 'Congrats on.' Ensure the final subject line does not exceed 10 words. Always use second-person pronouns. We are drafting a subject line for {FullName}, who works at {CompanyName}.”

‍

Here, dynamic variables such as {FullName} and {CompanyName} are utilized in the prompt to direct the AI to the specific data it needs from the Clay table. These dynamic variables enable the reuse of data across various datasets.

There are essentially two methods for generating AI snippets for personalized emails. The first and simpler method is outlined above. In this approach, we instruct the AI to generate a sentence by specifying how it should begin, what its purpose is (e.g., summarizing, commenting, praising), and its length. This method is best suited for short paragraphs containing no more than 15 words. Below are some examples:

‍

‍

To create a more advanced personalization snippet, we utilize the second method: designing an email template and incorporating placeholders (dynamic variables) enhanced with AI-generated snippets within the content.

‍

‍

We will then create various snippets for each section of our email, following the method outlined above. This includes the subject line, opening sentence, subsequent sentences, and even a PS at the end, if applicable. By leveraging AI formulas, we will combine all these elements into one cohesive email, customized for each prospect. It’s hard to believe that both emails below originated from the same template.

‍

‍

Note that the e-mails above start with a different opening. The first one starts with “I wanted to reach out because I saw your post…”, while the second one starts with “I saw the news about…”. Don’t panic. It’s still 100% automated. What we did here is a conditional formula, instructing the AI to select the appropriate opening sentence based on the information available for that specific prospect (such as a post, news, or case studies) using an “if-else” sequence statement.

‍

‍

Follow some examples. The copies below were provided by Jed Mahrle, from Practical Prospecting. If it wasn’t for him, I believe I would still be lost in the darkness, since he was the one that introduced me to Clay. In one raining morning, he woke-up with the right foot and decided to publish his “sales playbook” as a lead magnet in his website. A very simple pdf, without any formatting, but very, very informative. I spent an entire day thoroughly studying his strategies when I came across an email copy containing a dynamic variable that initially didn’t make any sense to me. I sad to myself, how on earth can he insert that information on an automated email? And there it was, in a footnote: “all that can be 100% automated with Clay”. Thanks Jed. You changed my life.

‍

‍

Now, let’s move up a notch, and see some real case studies.

‍

Case Studies

‍

Case # 1 - Lead Assassins

This example was provided by Ericson Dalusong, founder of the Lead Assassins agency and which is a great reference to me. This guy was a mentor at the Clay Bootcamp and has also worked in Clay itself. I think that gives a pretty good idea of his knowledge in GTM Automations. In this example, Ericson demonstrates some of his strategies for prospecting SaaS companies.

One particularly noteworthy approach involves scraping company websites to locate a "book a demo" section. His reasoning is that companies offering personalized demos typically sell high-ticket products, making them more aligned with his Ideal Customer Profile (ICP) and thus yielding a higher Lead Score.

Additionally, Ericson developed a highly sophisticated sequence of data inferences with Clay, to achieve his objective of sending personalized emails to prospects using data from their clients, competitors, and case studies. The sequence of actions he executed within the Clay table was meticulously designed to support this goal:

‍

1. (1) dentify all decision makers for those companies: founders and C-levels
2. (2) enrich and validate their e-mail data
3. (3) summarize the solution each company provides
4. (4) identify the ICP for each company
5. (5) identify the name of at least one of their clients
6. (6) extract any thought leadership that might be available on the internet (podcats, article, youtube video, etc)
7. (7) extract the name of a competitor for each company
8. (8) extract a summary of a case study provided by the company in their “case studies” website page
9. (9) And then, after crafting the e-mail copies, sending them to other platforms for cold outreach both via e-mail and LinkedIn.

‍

‍

Below follows the copy of the final e-mails that were sent in this campaign, where he uses most of the previously scraped data.

‍

‍

‍

Case # 2 – The Klin Agency

This example was provided by Patrick Spychalski, founder of the Klin Agency, a pioneer in GTM Engineering and a content creator for Clay itself. Patrick has worked on projects with prominent clients like ClickUp and Semrush and frequently shares valuable insights and videos on his LinkedIn profile.

In this instance, Patrick utilized a tool called Trigify.io, which enables the aggregation of data and signals from social media platforms, such as LinkedIn. When someone likes or comments on a LinkedIn post, Trigify.io can capture that interaction and automatically route it into a table using a webhook. This process collects details such as the individual’s first name, last name, engagement type, the specific post they interacted with, and the post’s URL.

Patrick then employed an AI prompt via ChatGPT to summarize the post, enriched company data with this information, and used an AI conditional formula to create industry-specific social proof. This approach allowed him to send tailored case studies to prospects based on their industry. For instance:

‍

• if the company operates in the “Moving” industry, they would receive a case study like: “See how Snoball helped New City Moving achieve 30+ additional moves in just 1 month.”

‍

• Similarly, businesses in the “Solar & Roofing” sector might receive: “Recently we helped ReNu Solar & Roofing get 80 referrals.”

‍

• The same logic applies to other industries, such as “Home Remodeling,” ensuring that each prospect receives relevant and personalized content.

‍

‍

The process doesn’t end there. He proceeded to locate business reviews and identified the name of an individual who had left a review for the company. He then summarized how the company assisted this reviewer and incorporated that information into a personalized message. The message was sent as part of a Smart Lead campaign, referencing the Google review and the reviewer. It stated:

‍

We noticed how you supported Ms. Denise Joens in ensuring safety for elderly individuals. Congratulations on your excellent work. We specialize in assisting companies like yours.

‍

The message concluded with a compelling social proof statement:

‍

Recently, we helped Zyntex secure 38 new referrals, resulting in over $74,000 in home remodeling services. Would you be open to discussing how we can assist you as well?

‍

‍

It’s just mind blowing how far we can go. You can keep moving up a notch indefinitely. The level of sophistication and personalization of what we can do with Clay in limitless. It depends just on your imagination. Continue reading until the end of this article, and you will gain a clear understanding of what I mean. You will be mesmerized.

‍

Case # 3 - The ClickUp Campaign for ClickUp Brain

ClickUp has created a campaign to reach out to managers at Fortune 500 companies, with the goal of getting them to adopt one of their new products internally: the ClickUp Brain.

‍

‍

Since the ClickUp brain is a low-cost SaaS, it may not require top management approval and, therefore, they did not need to target first tier decision makers. They used as signals: (1) company headcount growth, targeting companies with a growth above 10%; (2) companies that were acquired or made an acquisition; and (3) prospects that worked in departments that did administrative work related to back-ups.

And this was the personalized first sentences they used for each case:

‍

‍

We can only imagine the level of complexity involved in the formulas related to the dynamic variables {{JobTitleDocuments}} and {{timeSavedCalculation}}. Below is the continuation of the email content:

‍

‍

GTM Engineers: The sales job of the future

The Go-to-Market (GTM) Engineer is a role that originated with Clay. Interestingly, the initial GTM pioneers were former employees of Clay. The company has also published a detailed article that provides valuable insights into what a GTM Engineer entails:

‍

They'll develop deep expertise in modern sales technology - knowing when to use AI vs. human touch, creating automated data pipelines for prospect information, and continuously measuring and optimizing their systems for better performance.

‍

In my own words: it’s the sales professional whose only job is to generate more pipeline without increasing the headcount - and the headcount that is going to suffer is the SDR. Traditionally, the sales framework consists of four roles:

‍

• SDRs research prospects and send outbound emails;

• Account Executives (AEs) handle calls and close deals;

• Sales Engineers provide technical explanations of the product; and

• Revenue Operations (RevOps) ensure the CRM and tools function smoothly.

‍

A new role, the GTM Engineer, is emerging to optimize the Lead Generation process, enhancing SDR productivity. This means SDRs will no longer need to craft personalized emails. Cold Calls can also be optimized as Clay automates Lead Qualification, allowing SDRs to focus only on high-quality leads (scoring 8 or above).

Aamir Bajwa, founder of Corebits, highlighted in a LinkedIn post the cost comparison for Lead Generation between employing one GTM Engineer versus a team of five SDRs. His analysis shows that a GTM Engineer using an effective Clay outbound engine can reach 10,000 prospects per month. Considering the salary for one GTM Engineer and software expenses, the Cost per Acquisition (CPA) is USD $357. In contrast, a team of five SDRs results in a CPA of USD $600, leading to an estimated Customer Acquisition Cost that is USD $20,000 higher on a monthly basis.

Patrick Spychalski created a LinkedIn Post expressing the same ideas:

‍

Instead of hiring five SDRs to handle research, enrichment, and outreach, one person can now run that entire motion as a workflow. Use job change or hiring signals to surface leads, enrich with Clay + LinkedIn + Clearbit, pull in tech stack and intent data, and push to Apollo or Instantly for automated outreach. The pipeline refreshes daily - no manual work required. In 2025, the teams that win won’t be the ones who hire the most - they’ll be the ones who build the best systems.

‍

The role of SDRs (Sales Development Representatives) is not expected to become completely obsolete, but their numbers are likely to decline. As with many professions, the advancement of AI will render certain aspects of the SDR role unnecessary. In the near future, I anticipate a significant shift, with many SDRs transitioning into roles as GTM (Go-To-Market) Engineers. There will likely be an increased emphasis on “tech-savy” sales professionals rather than empathy and communication-based soft skills.

Now, you will be surprised with what I am about to say:

‍

As of the date this article was published, there were only 581 GTM Engineers globally (based on a LinkedIn Sales Navigator search). Yes, globally - not just in Silicon Valley, where, even then, that number would be remarkably low.

‍

It seems people did not yet understand what is going on.

One year from now, I plan to conduct this search again out of curiosity to see how these numbers evolve. If you'd like to make a prediction, feel free to share your estimate in the comments section.

‍

‍

Clay Beyond Sales

My wife is a university professor specializing in environmental engineering, with a focus on epidemiology. We are planning to live in Canada for a year starting next year (I am a Brazilian that loves snowboarding). She is currently seeking an academic professor in her field to potentially sponsor her for a postdoctoral project. Using Clay, I can compile a table of all university professors in Vancouver, categorize them by their areas of expertise, and conduct a search for keywords like “epidemiology” or “diseases” across their published works. This search can extend beyond LinkedIn and cover the entire internet. Once we identify professors with the highest "lead score," we can efficiently add them to a targeted email sequence and initiate contact.

This Clay article demonstrates how the platform can be used to create a coffee shops exploration guide. This guide can include locations, ratings, reviews, contact information, and even favorite menu items, with an optional column indicating the distance of each coffee shop from your home.

Recruiting and HR agencies are among the most significantly impacted by the new possibilities Clay offers. Recruiters can now identify candidates with unparalleled precision and implement “Candidate Qualification” processes similar to how sales teams qualify leads. For example, recruiters can analyze candidates’ professional experiences, duration of employment, and the companies they have worked for. If you need to fill a position for a very specific and rare knowledge, such as a roller coaster engineer for example, you can search for candidates based on that specific keyword. Additionally, if a CFO has recently transitioned to a new role at another company, you can extract information about their previous organization and send them a personalized message saying,: “Do you need to replace the CFO you recently lost? We can assist with that.”

Clay’s capabilities extend far beyond our wildest dreams. It allows users to enrich tables with multiple data points, including brand colors, company descriptions, website details, and more. This data can then be sent to external generative AI applications like Gamma to create AI-driven PowerPoint presentations in bulk through a fully automated process. The presentation links can be integrated back into the Clay table, enabling personalized outreach to each company with messages like: “We have some design ideas for you—can I send you a PDF sample?” The possibilities are astounding. Not only can emails be personalized, but entire videos, presentations, and even software can be customized.

‍

Conclusion

It seems that Clay is such a powerful tool that we do not yet know all we can do with it. It enables users to extract databases from the internet and leverage them for virtually any purpose. While I’ve outlined a few examples here, countless new applications and job opportunities will emerge as more people adopt this technology.

It is also very clear that tools like Clay, and AI in general, are already disrupting the market. Traditional sales tools that have not yet adopted AI are losing ground. Klue, a Canadian platform focused on sales intelligence, recently let go of 40% of its workforce due to mounting competition from AI-driven technologies.

‍

“A frequently asked question (FAQ) document Klue compiled for staff in the leadup to the layoffs that BetaKit has viewed indicated that the startup had failed to reach its topline revenue growth targets while seeing increasing customer churn this year. The FAQ noted that Klue has been facing difficulty competing against OpenAI’s ChatGPT offering. The FAQ stated specifically that Klue has been “losing more deals to ChatGPT (AI-disruption),” and is building a new product to address this and help the company “re-find product-market fit—with AI at the core.”

‍

It looks like that will also happen to us - sales professionals and SDRs – we either swim with the current or we will be out of the market.

‍

‍