There seem to be two ways to win in GTM right now: being exceptional at automation and being exceptional at relationships.

Both ends of this barbell are working. The space in between is falling flat.

But automation and relationships don’t need to be at odds with each other. Readers tell me they’re inspired by Clay as a rare company that strikes this delicate balance. (You also ranked Clay #2 on the list of ‘must try’ GTM tools this year.)

Clay makes infrastructure to launch and run self-learning GTM plays. They popularized the term GTM engineer in 2023 and have among the best in-house teams focused on automating GTM workflows. Meanwhile, Clay goes big on what’s hard to automate: storytelling, design, education, local meetups, a flagship IRL conference (Sculpt).

I turned to Osman Sheikhnureldin, an early employee and Clay’s head of GTM engineering, to explore exactly how Clay automates GTM without losing the human touch. We unpacked their 17 best automated GTM workflows that you can replicate. Keep reading to go deep into account scoring, routing, meeting prep, automated follow-ups, and much more.

Clay’s first goal is to prioritize where the sales team invests in relationship building. These workflows cover data enrichment, propensity scoring, territory planning, qualification, and routing.

Why this matters: Clay has 250,000 accounts in Salesforce and 850,000 users. Prioritizing the right accounts matters more than finding new ones.

Every account gets enriched with standard firmographic data points (headcount, headcount within specific departments and teams) along with custom AI research fields. I found Clay’s custom research fields to be particularly interesting since this data is so much easier to gather at scale now and becomes the foundation for personalization.

Clay tags each account with their primary personas and segments (i.e. who the account sells to) and whether they have a product-led growth (PLG) motion vs. sales-led vs. a hybrid of the two. Clay will also estimate the account’s total addressable market (TAM) since that’s a part of how teams use Clay.

This data gets translated into a fit score from 0 to 100 (low fit → best fit) and written to a Salesforce field. The inputs for the fit score: company size, business model (B2B vs. B2C), GTM motion (PLG, sales-led, hybrid), headcount in sales and operations, data warehouse, CRM, whether competitive products are in use, region, funding stage, and revenue trajectory.

Each data point has supporting evidence into why it matters. Osman investigated Clay’s top customers by ARR, for example, and found that these customers have 10%+ of their headcount in sales or GTM. A high sales headcount indicates that the account is actively investing in GTM, making it a great fit for Clay.

Clay then maps accounts into tiers based on the fit score (mentioned above) along with the seniority of the inbound contact. Each tier triggers a different set of actions:

The fit score is one of four inputs that roll up into an overall buying propensity score, which is what Clay uses to help a rep prioritize their accounts. This runs within a rep’s book every week and automatically flags the top 5 highest priority accounts to work.

The components of Clay’s propensity score:

Clay’s engagement score incorporates factors like website visits, social activity, product usage, and event and livestream attendance. All get smushed together as an indicator that they’re actively interested in Clay.

For a long time, Clay didn’t have sales territories and a rep’s book was exclusively inbound accounts. They moved to defined territories earlier this year.

At Clay, an account is added to a territory when its weighted Account Fit Score reaches above 60, it has no open opportunity, and there’s no assigned rep.

When a rep gets assigned an account, Clay proactively sources the most relevant contacts based on the personas that they know are involved in deals. It might sound trivial, but this can be a huge time saver for reps.

Clay tracks signals at both the account-level as well as at the contact-level (this is triggered by AEs/SDRs for their own contacts). Signals get pushed out as Slack alerts and live in Terra, a custom rep UI that Clay built. It brings together product usage, past job history, initiatives heard on calls, and why any prior deal was closed-lost.

The GTM engineering team has dedicated workflows for certain signals like a new hire or a job changer. Outreach drafts are generated automatically, but reps edit and approve them before they go out (human-in-the-loop).

If Clay spots a new executive hire at a target account, the play pulls the job posting that got that person hired or the job posting that’s closest to the job title. It looks up what that person was hired to do, which goes into the outbound copy.

If a Clay user or champion changes jobs, the play looks up how that person engaged with Clay in their last job (ex: product user, champion) and tailors the message based on that context.

Related: 4 GTM plays that combine first-party and third-party data

At this stage reps have their accounts, know the right contacts, and have an initial demo scheduled. The GTM workflows shift toward making sure reps are prepped with all the right context before each conversation and keeping stalled deals moving.

Three days before a demo call is scheduled to occur, Clay takes what the prospect submitted on the lead form and tailors a pre-demo email accordingly. (“We saw you’re interested in X. What else should I know before we talk?”)

This gets sent to the contact that came inbound along with everyone else added to the calendar invite. The AE reviews and approves the email before it’s sent.

At the start of each week, reps get a research brief about every account they’ll be meeting with. This is done for every meeting and the brief changes based on how many meetings the rep has had with the account.

If it’s the first meeting, the brief focuses on what the prospect put on the form, who they are, and who they sell to. They use Clay to figure out the top 3 use cases for the account (this also draws from the custom data about whether the prospect has a product-led, sales-led, or hybrid GTM motion). The brief also researches the people at the account, whether they’ve used Clay before, whether they’re already a product user, and whether they’ve had access to Clay at a prior job.

If it’s a second meeting (or later), the brief recaps past calls, open next steps, and new stakeholders.

For deals that have been stuck in a specific stage for a while, Clay will pull additional contacts to reach out to. This includes the most senior GTM contact they can find at the account.

Clay has found that a sales leader to sales leader connection is one of their most effective plays for getting these deals unstuck. Reps can click a button to queue up an email that would go out from their Head of Sales (Becca). It lands in Becca’s Gmail drafts for her to edit and send manually.

Clay’s Majors team works complex Fortune 500 deals. These span multiple teams and business units that might not talk to each other. A big pain here is simply figuring out who reports to who and what each team’s mandate is.

They’ve built an account agent that pulls from both internal activity (i.e. call transcripts and email activity that mention reporting structures) and external job posts (i.e. a Claygent finds current and archived job postings since that’s a rich source of information about a company’s reporting structure). This org chart workflow is live, meaning that it stays up to date as there are new learnings from calls and emails.

This next phase focuses on capturing what happened on calls, turning it into follow-ups in the rep’s voice, and proactively handling potential objections that could block a deal.

Once a rep has a call, they get a follow-up email that’s sitting in their email drafts. Follow-ups aren’t novel. What’s special is that Clay writes them in the rep’s own voice.

To do that Clay’s GTM engineers pull into Snowflake all the non-automated things that the rep has either sent or said on calls. That gets turned into a prompt that models how the rep naturally writes (common words or phrases, tone of voice, sentence length, etc.).

In addition to the follow-up email, the rep gets a form pre-filled with what to update in CRM based on the call. The rep just has to confirm the form. This speeds up CRM hygiene. It’s similar to features seen in Gong or Momentum, but more customizable to Clay’s sales process.

This workflow is probably the hardest to build (and Osman admitted that it’s in beta). Clay has pulled together information buried in its closed won and lost deals to compile a list of things that the best reps do to help them win. Other reps probably don’t know about these since they rarely get documented.

The workflow extracts from calls, Slack and email all the things that need to happen to move the deal a stage ahead. It’s a mix of to-do items whether they’re on the prospect’s side or on Clay’s side. For example, the prospect might owe a list of contacts to get on a trial account or an introduction to procurement. Or the Clay AE might owe a comparison doc of Clay vs. a different vendor or an intro to an expert on Salesforce integrations with Clay.

Importantly, it also identifies the most likely objections on the deal and how to proactively respond to those objections. Today this lives outside the Clay product, but the team has plans to build this in.

One example of a potential objection: there’s rarely just one decision maker in an enterprise deal and the sales team doesn’t have a relationship with every stakeholder who matters. Clay’s multi-threading workflow pulls who was on prior calls and who’s engaged, then finds the next people to multi-thread with (i.e. not just who’s already in the CRM).

Historically, post-sale hasn’t had as much of a GTM engineering focus at Clay. That’s changing quickly with workflows for the sales to CX hand-off and account health scoring.

When a deal closes, a hand-off gets automatically created via Clay with all the context on the deal so CSMs have what they need to know. It surfaces which use cases the account wants to do first and then what else they could do after that.

There’s also a workflow the CSM can fire that auto-generates a kick-off deck specific to the account. These decks pull account data from Salesforce, calls from Gong, and credit usage from Snowflake. (Clay runs a similar workflow to automate QBR decks, which they estimate saves 90 hours per CSM each quarter.)

As the customer uses Clay, GTM engineering built a workflow to tag accounts with a score to identify who’s at the highest risk of churn. The score is built on product data (i.e. credit burn vs. what was purchased), what’s said on calls and in shared Slack channels (every enterprise account has one), and third-party signals on the account itself (ex: did they have a round of layoffs, did the champion on the deal leave).

Three things stood out from Clay’s process that I think are worth replicating:

Clay shows you don't have to pick an end of the barbell. Automate everything around the relationship so your people can spend their time on it.