Clay for RevOps 2026: workflows, scoring, and what it costs

How do you run Clay across a whole revenue engine, and what does it cost to run in 2026? This guide answers both. Clay is the enrichment, research, scoring and signal layer under a RevOps stack, and since Clay repriced on 11 March 2026 the cost question has a new answer: you now budget two separate pools, and the one most teams exhaust first is not the one they watch.

Updated September 2026: added Clay's March 2026 repricing and the data credits versus actions split, sourced or replaced every statistic on the page, and rebuilt the scoring section around a rubric you can copy.

The problem this solves is well measured. Salesforce's State of Sales research puts the share of a rep's week actually spent selling at well under half, with the rest going to research, admin and internal process (Salesforce); the reported figure moves between roughly 28% and 40% across editions and summaries of that research, so treat it as a band rather than a number. Contact data underneath that work decays at about 2.1% a month, which compounds to roughly 22.5% a year, a benchmark that traces to MarketingSherpa and is reproduced in HubSpot's database decay model. Roughly a quarter of a B2B database goes materially wrong within twelve months whether or not anyone touches it.

That is the gap Clay is bought to close. If you want the full case for Clay as the centre of a modern GTM stack, including the pricing history, see our deep dive on the broader case for Clay as a GTM platform.

At Ziel Lab, we build Reasoning Agents, not simple automations. We treat Clay as the data layer for agentic RevOps, where AI does not just inform decisions but executes them. This guide maps each Clay capability to a revenue outcome, gives the implementation blueprint, and prices it.

What is Clay and why does it matter for RevOps?

Clay is a data enrichment and workflow automation platform that brings AI agents, multi-source enrichment, and intent signals together in one place. But calling it just an enrichment tool undersells its capabilities significantly.

At its core, Clay functions as a GTM development environment. Rather than forcing you to choose one data provider, Clay gives you access to over 150 data sources through a single platform (Clay's own figure). Instead of building rigid automation sequences, Clay enables intelligent workflows that reason about your data and take contextual action.

The platform works like a spreadsheet that thinks. You import your leads or accounts, add enrichment columns that pull data from dozens of providers, apply AI analysis to identify patterns, and push the results to your CRM or outreach tools.

What makes Clay different from legacy enrichment tools

Waterfall enrichment is Clay's signature approach. Query multiple data providers in a prioritized sequence. If Provider A does not have the email you need, Clay automatically falls to Provider B, then C, continuing until it finds the data. You pay for successful searches.

One correction worth making, because this page carried the error for a year and other write-ups repeat it. The often-quoted jump from the low 40% range to the high 80% range in inbound enrichment coverage is OpenAI's number, not Anthropic's. Anthropic's published result is a separate one: roughly triple the enrichment coverage of the single provider it replaced. Both are Clay's own customer stories, written by the vendor, and neither is audited. Read them as an indication that waterfalling beats single-sourcing on coverage, which is the mechanical claim and a safe one, not as a coverage rate you should expect to hit.

Live web scraping means Clay enriches from real-time sources including websites, job boards, social profiles, and news articles. You are not limited to static databases that were updated months ago.

AI-native research through Claygent, Clay's AI research agent, which the company says has passed one billion runs. Unlike bolted-on AI features, Claygent reasons about your research needs in natural language. Ask it to find a company's recent product launches, identify key competitors, or extract personalization angles from a prospect's LinkedIn activity.

Positioning Clay in your RevOps stack

Think of your stack in three layers:

• Clay = the GTM data layer (enrichment, research, signal capture)
• HubSpot or Salesforce = the system of record (pipeline management, reporting, automation triggers)
• n8n or similar orchestration = the reasoning layer that connects signals to actions

This architecture differs fundamentally from traditional setups where you have a CRM that stores information, a single-provider enrichment tool, and basic if-then automation. Clay actively discovers and enriches data. It waterfalls through multiple providers. And when combined with reasoning-capable orchestration like what Ziel Lab engineers with n8n, it creates workflows that adapt to context rather than following rigid scripts.

Key takeaway: Clay is not a replacement for your CRM. It is the intelligence layer that makes your CRM actually useful by ensuring every record is complete, enriched, and actionable.

Core Clay features that power RevOps

Every Clay capability maps directly to RevOps outcomes. Here is how each feature translates to metrics that matter for your revenue team.

Data sources and waterfall enrichment

Clay pulls company and contact data from multiple vendors including its own data graph, LinkedIn-style sources, Apollo-like databases, and live website scraping. The waterfall enrichment model queries providers in your defined priority order until one returns the data you need.

RevOps impact:

• ICP precision and list quality: Higher match rates mean better reply rates and more meetings booked
• Cleaner CRM data: Accurate records enable proper routing, reliable forecasting, and trustworthy reporting
• Reduced manual research time: SDRs spend time selling instead of Googling prospects

IntroCRM, a fractional sales agency, reports cutting its prospecting data budget by 65% after replacing a single expensive data broker with waterfall enrichment (Clay case study). That is a vendor-published number from a customer with an incentive to be quoted, so weigh it accordingly. The mechanism behind it is the part that generalizes: paying only on a successful return, rather than for access to a database whether or not it holds your record, is a different cost structure, and it favours you most when your list is one a single provider covers badly.

Implementation blueprint:

• Build an ICP template in Clay starting from domains or LinkedIn search
• Enrich with revenue, employee count, industry, geography, and tech stack
• Filter to ICP tiers using computed columns or AI scoring
• Push enriched, scored accounts into HubSpot via Clay's native integration or through n8n for advanced logic
• Set up n8n workflows to monitor new contacts in HubSpot, send them to Clay for enrichment, write enriched fields back, and trigger routing workflows

AI-powered research with Claygent

Claygent performs research tasks that previously required human effort at scale. The AI can summarize websites and about pages, extract buying triggers like expansion signals or hiring patterns, identify relevant value propositions from a target's site, auto-generate company one-liners, and execute multi-step research workflows.

RevOps impact:

• Context-rich CRM records: Better qualification and faster discovery calls
• Scalable personalization: Relevant outreach at sequence start without manual research
• Shorter ramp time: New reps use AI-generated account overviews instead of spending hours researching

Implementation blueprint:

• For each new account in HubSpot, trigger an n8n workflow that sends the domain to Clay
• Clay scrapes the website, careers page, blog, and product pages
• Claygent summarizes what the company does, who they sell to, and why they might need your product
• Write the summary and ICP fit reasoning back to HubSpot as custom fields
• Use the AI summary in HubSpot record sidebars for AEs, as input for AI-generated first-touch emails, and in qualification playbooks for SDRs

AI personalization and message generation

Clay generates row-level personalized content including email first lines, complete emails, LinkedIn openers, and call scripts. It uses person role, company context, website snippets, job post details, and your value prop templates to create unique, relevant messages for each prospect.

RevOps impact:

• SDR productivity: One rep can prepare personalization across a list that would otherwise take a research day, because the research runs per row instead of per person
• Improved deliverability: Unique, non-template content avoids the pattern matching that catches mail-merged sequences
• Messaging consistency: Personalization aligns with RevOps strategy and ICP positioning

On the size of the lift, the most credible number we found is not Clay's. Unify's analysis of more than 25 million outbound sends puts AI-personalized emails at 57% more replies than generic templates, and is explicit that the lift only holds when the personalization is grounded in real research rather than a merged first name (Unify). Clay's own customer story on this is ServiceBell, 30 meetings booked from about an hour of work, again vendor-published and unaudited.

The honest read: personalization at this depth is worth a meaningful multiple on reply rate, not an order of magnitude, and the multiple collapses if the AI column is writing from thin research. Budget for the research, not just the writing.

Implementation blueprint:

• Define messaging frameworks with your RevOps team mapping value props to ICP segments and personas
• In Clay, add AI columns to generate personalized first lines and email bodies using those frameworks combined with enriched signals
• Push final email copy and variables into your sequencer (Instantly, Outreach) or back to HubSpot sequences
• Use n8n to trigger Clay AI personalization when a prospect moves to Ready for Outreach stage
• Add an AI QA agent to check compliance and tone before sending

Lead and account scoring

Clay calculates scores using computed columns based on firmographic fit (size, industry, region), technographic fit, and signals like hiring for relevant roles or tech adoption. AI-based ICP fit scoring combines multiple inputs into a 0 to 100 score with prioritization logic for Tier 1, 2, and 3 accounts.

RevOps impact:

• Aligned definitions: MQL, SQL, and ICP mean the same thing across sales and marketing
• Better lead routing: Best reps get best leads with proper territory alignment
• Higher pipeline quality: More accurate forecasting and better quota attainment distribution

Implementation blueprint:

• Translate your ICP definition into a Clay scoring rubric, then calculate it in a computed column
• Send the result to HubSpot as two fields, ICP Score and ICP Tier, and keep the reason string as a third
• In HubSpot, route Tier 1 to named AEs instantly, Tier 2 to the SDR pool, Tier 3 to nurture
• Build an n8n reasoning layer that watches high-score account behavior (website visits, email opens) and decides whether to escalate, send tailored content, or wait

The rubric we actually run

Most guides stop at "build a scoring rubric" without saying what goes in one. This is the shape we start from on Clay plus HubSpot engagements, before tuning it to the client's data. It is deliberately blunt: 100 points, three buckets, one tie-breaker.

Mapping revenue bands to scores is the part people ask about most, and the mistake is almost always making it linear. Fit is not linear. A band either sits inside the range where your product is bought or it does not, and the shoulders on either side are worth partial credit, not proportional credit:

• Inside the core band: full firmographic points
• One band either side: about 60% of them
• Two bands out: about 20%
• Beyond that: zero, and no amount of signal should rescue it

That last clause is the one that earns its keep. Without a hard floor, a 40-person company with three relevant job posts outscores a perfect-fit account that happens to be quiet this quarter, and your AEs learn to ignore the score within a month. Cap signals so they can promote a good-fit account, never manufacture one.

Two more things we have learned the expensive way. Write the reason string, not just the number: a score with no explanation gets argued with, a score that says "core revenue band, runs HubSpot, hiring two RevOps roles" gets acted on. And re-score on a schedule rather than once at capture, because the firmographic half is exactly the data that decays at 2.1% a month.

Signal capture for timing-based outreach

Clay monitors job boards, company careers pages, LinkedIn changes, funding news, and tech stack updates through enrichments and web scraping. You can refresh tables on schedule (daily or weekly) to pick up new signals and use AI to interpret whether signals are relevant to your value proposition.

RevOps impact:

• Operationalized signal-based selling: Move from static lists to dynamic, intent-driven outreach
• Better timing: Contact prospects right after a funding round or when they post jobs for your champion role
• Stronger narratives: Multi-touch campaigns around events and milestones resonate more

Implementation blueprint:

• Build a Clay table of ICP companies seeded from HubSpot, Apollo, or CSV uploads
• Add enrichments for latest funding round (amount and date), active job postings mentioning relevant keywords (RevOps, HubSpot, AI), and technology adoption signals
• Schedule automatic refresh of these signals weekly or daily
• Configure n8n to detect new or changed signals on each refresh
• Trigger actions: create a task in HubSpot for the AE or SDR, add the contact to a signal-based outbound sequence, or kick off an AI agent to generate contextual outreach

Multi-source data unification and de-duplication

Clay combines multiple sources like Apollo exports, LinkedIn searches, manually uploaded lists, and internal datasets into unified tables. It deduplicates by domain, email, or LinkedIn URL and lets you define field priority rules when sources conflict.

RevOps impact:

• Single source of truth: Clean GTM data before it enters your CRM
• Reduced duplication: Fewer split ownership issues and routing errors in HubSpot or Salesforce
• Better analytics: Accurate attribution and account-based reporting

Implementation blueprint:

• Build a Data Master Table in Clay with inputs from product signups, event lists, Apollo exports, and webinar attendees
• Normalize and dedupe using Clay's matching capabilities
• Use Clay as a pre-CRM hygiene layer, only pushing clean, enriched, deduped records into HubSpot
• Periodically export CRM records to Clay via n8n for cleanup and re-enrichment, then write back corrected and standardized fields

Key takeaway: Every Clay feature connects directly to revenue metrics. Waterfall enrichment improves coverage and reduces costs. AI research accelerates qualification. Personalization increases reply rates. Scoring improves routing. Signals improve timing. Unification improves data quality.

RevOps strategies you can implement with Clay

Features become valuable only when they drive strategic outcomes. Here are five concrete strategies you can implement immediately using Clay combined with HubSpot and n8n.

Strategy 1: Build a signal-based ICP engine

Goal: Move from static ICP definitions to dynamic, signal-based scoring that updates automatically.

Static ICP lists decay quickly. Companies grow, shrink, adopt new technologies, and enter buying cycles. A signal-based ICP engine keeps your targeting current and focuses sales effort on accounts showing active buying intent.

How Clay does it:

• Base table of all target accounts (your TAM or near-TAM)
• Enrich with tech stack, revenue, employee count, region, funding, and job postings
• AI column generates ICP Fit Score (0 to 100) and Reason for Fit
• Weekly refresh captures changes in signals

n8n orchestration:

If score exceeds 80 and a new relevant signal appears, create a task and outbound sequence trigger in HubSpot. This ensures high-fit, in-market accounts get immediate attention while lower-priority accounts stay in nurture.

Bottom-line impact: Better focus of sales effort on in-market accounts. Higher conversion rates and pipeline per rep.

Strategy 2: Automated enrichment and routing for inbound leads

Goal: Every inbound lead gets fully enriched and routed according to ICP and intent automatically, within seconds of form submission.

Speed to lead matters, and the effect is steep rather than gradual: the widely cited finding is that contact and qualification rates fall away sharply once the response gap moves from minutes into hours. The original study behind that claim is old enough now that we would not put a precise multiple on it in 2026, and we have not found a recent replication we trust, so treat the direction as established and the magnitude as unverified. What is not in doubt is the mechanism you control: manual enrichment and routing put the delay there in the first place.

Implementation flow:

• Trigger: New HubSpot contact from form submission
• n8n extracts domain from email and sends to Clay
• Clay enriches company and contact, calculates ICP Tier and Score, generates AI summary of why they might be interested
• n8n updates HubSpot with enriched data
• Routing logic: Tier 1 triggers deal creation, AE assignment, and internal Slack alert with Clay summary. Tier 2 sends to SDR queue with AI-personalized first-touch email. Tier 3 adds to nurture workflow.

Bottom-line impact: Faster speed to lead. Better lead-to-opportunity conversion. Higher AE efficiency with fewer low-fit meetings.

Strategy 3: Multi-touch outbound with AI personalization at scale

Goal: Use Clay to power personalization at volume for outbound sequences across email, LinkedIn, and phone.

Generic outreach gets ignored. But manual personalization does not scale. Clay bridges this gap by generating unique, contextual messages for every prospect based on real data.

Implementation flow:

• Build a Clay table of prospects from LinkedIn search, Apollo, or event attendees
• Enrich with company signals: funding news, content topics, tech stack, hiring patterns
• AI generates personalized first lines referencing recent company events, persona-specific value propositions, and custom call script bullet points
• Push to sequencer (Instantly, Outreach) for emails and HubSpot tasks for LinkedIn and calls
• Optionally use Ziel Lab's Surround Sound Architecture to orchestrate across email, LinkedIn, SMS, and automated direct mail

Bottom-line impact: A meaningful lift on reply rate against generic templates, on the order of the 57% Unify measured across 25 million sends, provided the personalization is grounded in real research. More meetings per SDR without increasing workload.

Strategy 4: Account research packets for AEs

Goal: Before first calls, reps have deep context on the account without spending hours on manual research.

Discovery calls suffer when reps lack context. They ask generic questions. They miss obvious pain points. They fail to connect their solution to the prospect's actual situation.

Implementation flow:

• For each target account, Clay scrapes website, product pages, pricing (if public), careers to see hiring focus, and blog for key themes
• Claygent summarizes: What they do. How they make money. Possible pains related to your solution. Questions to ask on the first call.
• When an opportunity is created or moves to a specific stage, n8n triggers an update of Clay research
• Save AI summary to HubSpot as a note or custom field accessible in the deal record

Bottom-line impact: Shorter discovery time. Better call preparation. Higher close rates from more relevant conversations.

Strategy 5: Continuous CRM cleanup and enrichment

Goal: Use Clay as an ongoing data hygiene and enrichment layer for HubSpot or Salesforce.

Contact data decays at about 2.1% a month, compounding to roughly 22.5% a year on the MarketingSherpa benchmark reproduced in HubSpot's decay model. Job titles change, people leave, domains move, emails bounce. Note that this benchmark is cross-industry: in tech and SaaS, where role changes and funding-driven headcount swings run faster, the real rate is higher, so read 22.5% as a floor rather than an estimate. Without continuous maintenance your database degrades on its own, at a rate you can plan around.

Implementation flow:

• n8n runs a nightly job to extract newly created or modified accounts and contacts from HubSpot
• Send records to Clay for de-duplication, email verification, and enrichment
• Clay uses AI to normalize job titles and industry tags, marking bounced or invalid emails
• Write corrected records back to HubSpot
• Flag significant changes (person left company, company acquired) for human review

Bottom-line impact: Improved data quality. Better segmentation and reporting. Less manual cleanup by RevOps.

Key takeaway: These strategies transform Clay from a point solution into a strategic RevOps capability. Start with one strategy, measure results, and expand based on what drives the most impact for your team.

Clay plus HubSpot plus n8n: building the RevOps reasoning engine

The real power of Clay emerges when you integrate it with your CRM and orchestration layer. This creates what we at Ziel Lab call a Reasoning Engine, a system that does not just follow rules but interprets context and makes intelligent decisions.

The architecture

HubSpot is your system of record. It manages pipeline stages, stores customer data, runs automation workflows, and generates reports. It is where your sales and marketing teams live day to day.

Clay is your enrichment and AI research layer. It gathers data from 150 plus sources, runs AI analysis, and prepares records for action. It is the intelligence that makes HubSpot data useful.

n8n is the reasoning layer. Unlike basic if-then automation, n8n workflows built by Ziel Lab use LLMs to analyze context, make decisions, and adapt to the data in real-time. This is what separates simple automation from true intelligent agents.

Integration methods

Clay's native HubSpot integration allows you to push and pull contacts, companies, and deals directly. This works well for straightforward enrichment workflows where you want to sync data bidirectionally.

Webhook-based integration via n8n enables advanced logic. When you need multi-step reasoning (Is this a VIP lead? Is data missing? What is the best next action?), n8n provides the orchestration layer that Clay's native automation cannot match.

CSV import and export remains useful for legacy tools or one-time data projects where real-time sync is not required.

Example full flow

• Website form submission creates a new contact in HubSpot

• n8n detects the new contact via webhook

• n8n sends the contact's domain to Clay for company plus person enrichment

• Clay waterfalls through providers, calculates ICP score, runs Claygent to generate an AI summary

• n8n receives the enriched data and updates HubSpot

• n8n evaluates the ICP Tier:Tier 1: Creates deal, assigns AE, sends Slack notification with AI summary

• Tier 2: Sends to SDR queue, triggers Clay AI personalization, enrolls in outbound sequence

• Tier 3: Adds to nurture workflow

• If any data is missing, n8n routes the contact to a Clay table for manual research flagging

This is not linear if-then automation. The system interprets signals, researches missing information, and routes based on context. It adapts when variables change rather than breaking.

Key takeaway: The combination of Clay (data) plus HubSpot (record) plus n8n (reasoning) creates a RevOps engine that runs without a person in the middle of every step. To be precise about what is and is not claimed here: OpenAI, Anthropic and Notion are published Clay customers, which is Clay's claim and a checkable one. Their orchestration and CRM choices are not public, and this is our architecture rather than theirs.

Measuring the impact of AI-powered RevOps with Clay

Every investment needs measurement. Here is how to connect Clay usage to the metrics that matter for your revenue team.

Key RevOps metrics influenced by Clay

Lead-to-opportunity conversion rate: Better enrichment and routing means higher conversion from MQL to SQL to Opportunity.

Meeting booked rate: Better personalization and signal timing translates to more meetings from the same outreach volume.

Reply rate on outbound: AI-personalized messaging earns higher engagement than generic templates.

SDR productivity, meetings per rep: Automation multiplies output without increasing headcount. Notion reports using Clay to research and vet companies applying to its startup program, taking the auto-approval rate to roughly 40% with no manual review (Clay customer story), which is a vendor-published figure for a qualification workflow rather than an outbound one.

Time to first meeting: Faster enrichment and routing shortens the gap between lead capture and first conversation.

Pipeline coverage and quality: Better ICP scoring creates more qualified pipeline with higher close probability.

Win rate: Better discovery and targeting from AI research leads to higher close rates.

Customer acquisition cost: Efficiency gains across the funnel reduce the cost of acquiring each customer.

ROI framework

Calculate your return on Clay investment with this simple framework:

Time saved:

• Hours per week spent on manual research times hourly cost of that labor
• Example: 10 hours per week times $50 per hour equals $500 per week or $26,000 per year

Additional pipeline:

• Incremental meetings booked due to better targeting and personalization times average deal value times close rate
• Example: 20 additional meetings per month times $50,000 ACV times 20% close rate equals $200,000 additional pipeline per month

Cost reduction:

• Savings from consolidating multiple data subscriptions into Clay's credit model
• IntroCRM cut their prospecting data budget by 65%

Total value: Sum of time saved plus additional pipeline plus cost reduction compared against Clay subscription plus implementation cost.

Attribution best practice

Use HubSpot campaign and source tracking to attribute pipeline to Clay-enriched leads. Create a custom property in HubSpot indicating whether a contact was enriched by Clay. This allows you to compare conversion rates, deal velocity, and win rates between Clay-enriched and non-enriched cohorts.

Key takeaway: The ROI of Clay compounds across multiple dimensions: time saved, pipeline generated, and costs reduced. Track each dimension separately to understand where Clay delivers the most value for your specific operation.

The future of RevOps: GTM engineering and reasoning agents

We are witnessing a fundamental shift in how companies approach revenue operations. The discipline of GTM engineering, which Clay named in 2023, treats go-to-market as an engineering problem rather than a series of manual processes.

From automation to agents

Traditional automation follows linear, rule-based logic. If this happens, then do that. When variables change or exceptions occur, the automation breaks.

Agentic AI is different. It is contextual, adaptive, and capable of multi-step reasoning. An agent does not just execute a script. It analyzes the situation, considers options, and chooses the appropriate action. When something unexpected happens, it adapts rather than failing.

At Ziel Lab, our agents follow logic, not scripts. We build workflows that:

• Research a prospect's LinkedIn before emailing
• Read contracts and update deal stages based on content
• Qualify leads based on nuance, not just keyword matching
• Self-heal when APIs update or data formats change

Clay as the data layer for agentic RevOps

Clay provides the real-time data, enrichment, and AI research that agents need to make intelligent decisions. Without accurate, complete data, agents cannot reason effectively. Clay solves this problem by ensuring every record in your system is enriched with the context agents need.

n8n orchestrates the reasoning logic, determining what actions to take based on the data Clay provides. HubSpot remains the system of record, where outcomes are tracked and reported.

This architecture represents the future of RevOps: intelligent systems that research, reason, and execute with minimal human latency. The teams closest to their data are closest to new revenue.

The GTM engineer role

The role is now a real hiring category rather than a label. Clay's own April 2026 hiring guide put roughly 100 GTM engineer listings going live per month, and aggregators tracking the title counted more than 3,000 open roles across major job boards by March 2026. Both counts are soft in the same direction: the same job gets posted as GTM engineer, growth engineer, RevOps engineer or marketing engineer depending on the company, so any single figure undercounts and no two trackers agree. The trend is unambiguous even if the number is not. These professionals build revenue engines using AI and automation, measuring success by metrics like meetings booked and hours saved.

The work progresses through three stages:

• Data foundation: Keep CRM records clean and trustworthy through automated enrichment and deduplication
• Data modeling: Collect unique data points that predict purchase intent or churn risk
• Data activation: Deploy those data points in revenue-generating workflows

Key takeaway: The shift from automation to agents is happening now. Companies that build reasoning capabilities into their RevOps stack will outperform those relying on manual processes and rigid automation.

Getting started with Clay for RevOps: a practical roadmap

Implementing Clay successfully requires a structured approach. Here is your roadmap from initial setup to scaled operations.

Step 1: Define your ICP and scoring criteria

Before touching Clay, document your Ideal Customer Profile clearly. Include firmographic qualifiers (company size, industry, geography), technographic signals (tech stack indicators), and behavioral triggers (hiring patterns, funding events). This definition becomes the foundation for everything you build in Clay.

Step 2: Build your first Clay table

Start with a focused use case rather than trying to boil the ocean. Good starter projects include:

• Enrich inbound leads with company and contact data
• Build an outbound list for one vertical or territory
• Clean and dedupe a specific segment of your CRM

Import a small batch (100 to 500 records) and test your enrichment logic before scaling.

Step 3: Connect Clay to HubSpot

Use Clay's native HubSpot integration for straightforward sync or configure n8n for advanced logic. Test the data flow with a small batch to ensure fields map correctly and nothing breaks downstream.

Step 4: Add AI enrichments incrementally

Start with company summaries from Claygent. Once those work reliably, add personalization (first lines, email bodies). Then layer on scoring and signal detection. Each addition should be tested before proceeding.

Step 5: Build the orchestration layer

Create n8n workflows that route, score, and trigger actions based on Clay data. Start simple: when Clay enrichment completes, update HubSpot and create a task. Then add complexity: evaluate ICP tier, choose routing, trigger sequences.

Step 6: Measure and iterate

Track the metrics outlined earlier. Which enrichments drive the most value? Where are you wasting credits? What scoring weights produce the best conversion rates? Refine based on data, not assumptions.

What Clay costs to run in 2026, and where the action ceiling bites

This is the question we get asked most and the one this guide used to skip entirely. Clay overhauled its pricing on 11 March 2026, and the change matters more for RevOps teams than the headline prices do.

What changed

The Starter, Explorer and Pro tiers are gone for new customers. Two self-serve plans replaced them, Launch at $185 per month and Growth at $495 per month, with Enterprise on quote (Landbase, Cleanlist, Salesmotion). Existing customers on legacy tiers keep their pricing.

The structural change is the one to plan around. Billing now draws on two separate pools instead of one balance:

Sources: Landbase, Salesmotion, Cleanlist.

The arithmetic nobody does before they build

An action is spent per step, per row, and it does not matter what the step is. A formula costs the same as an AI prompt costs the same as a webhook. So a five-step workflow run against 1,000 records costs about 5,000 actions before a single data credit is touched.

Put Launch's 15,000 actions against that and the ceiling arrives fast. A moderately complex prospecting workflow gets you somewhere around 3,000 to 3,500 leads a month, and a ten-step table on 1,000 rows can consume most of a monthly allocation in one run (Bitscale's breakdown of the action limits).

This is what teams mean when they say they hit the ceiling on every large enrichment job. It is usually not the data spend. It is the orchestration around it, and under the old single-balance model it was invisible because it came out of the same pot.

The counterweight, because the discourse is skewed

The loudest voices on this change are agencies, high-volume power users, and vendors selling alternatives, all of whom are genuinely affected and none of whom are the median customer. Clay's position is that most customers will never approach the new limits, and a voice-of-customer analysis of the backlash makes the same point: the teams quietly modelling the change and finding it fine do not write posts about it (GTM Daily). One data point on the other side, from the same body of review analysis: around 42% of negative Clay reviews mention credit burn, which was true before this repricing too.

Our read, from running this stack: if you enrich a few thousand records a month and keep workflows tight, the two-pool model is a non-event and the marketplace data cuts that came with it may leave you better off. If you run wide tables with many steps, or you run them on behalf of clients, model your action spend before you commit to a plan, because that is the pool that will stop you.

How to spend less of both pools

The old advice on this page was written for a single credit balance and no longer maps cleanly. Split it by pool:

Spend fewer actions:

• Cut steps before you cut rows. Every column you can fold into one formula or one AI call saves an action per row, forever
• Filter first, enrich second. Run cheap conditional logic to disqualify rows before the expensive part of the table touches them
• Do not sync what you do not need. Every CRM write is an action, so push the fields that drive routing and leave the rest in Clay
• Watch scheduled refreshes. A weekly refresh on a 5,000-row table is a standing monthly cost, not a one-off

Spend fewer data credits:

• Order cheaper providers before expensive ones in the waterfall when quality is comparable, and stop the waterfall at the depth where returns actually fall off
• Deduplicate before enriching, not after
• Test on 50 rows before running 5,000

Watch both: check consumption weekly, per pool. A table that looks cheap on enrichment can still be expensive to run, and the reverse is just as common. If you want the pricing history and the plan-by-plan comparison in more depth, our Clay deep dive covers it, and the enrichment and routing guide works through the free-trial case specifically.

Ready to speed up your implementation?

If you want to shortcut this process and build a RevOps reasoning engine tailored to your stack, Ziel Lab can help. We specialize in Clay plus HubSpot plus n8n implementations for revenue teams ready to move beyond manual data work. Our approach combines CRM architecture expertise, intelligent workflow engineering, and signal-based outbound orchestration to build systems that research, reason, and execute autonomously.

Conclusion: from manual data work to intelligent revenue engineering

The evidence is clear. RevOps is evolving from manual data wrangling to intelligent, automated systems. Clay is a critical enabler of this shift, providing the enrichment, AI research, and signal capture layer that powers modern go-to-market operations.

OpenAI, Anthropic, Notion, Rippling and Verkada are among Clay's published customers, which tells you the tool clears a serious bar. It does not tell you what any of them built with it, and nobody outside those companies knows.

The architecture is straightforward: Clay for data, HubSpot for records, n8n for reasoning. Together, they create a RevOps reasoning engine that operates at machine speed with human-level intelligence.

Key takeaways from this guide:

• Every Clay feature maps to specific RevOps outcomes and metrics
• Implementation should be iterative, starting with one use case and expanding based on results
• Since March 2026 you budget two pools, and actions are the one that stops most teams
• The combination of enrichment plus CRM plus orchestration creates compounding value

Your GTM motion is probably not understaffed. It is under-engineered.

At Ziel Lab, we do not just set up tools. We engineer revenue systems. If you are ready to build a RevOps stack that reasons, researches, and executes, let us talk.

How we sourced this update

One limitation worth stating plainly rather than hiding. Clay's own documentation and several of the pricing publications cited above are not reachable from the environment this update was written in, so the plan prices, the credit and action allocations, and the per-row action arithmetic were established from multiple independent write-ups that agree with each other, and are attributed to the publication that carried them. Where sources disagreed, as they do on the share of a rep's week spent selling, this guide gives the range and says so rather than picking the tidier number. Vendor case studies are labelled as vendor case studies throughout. Verify the plan figures against your own Clay account before you budget against them.

‍