Which tools let a sales manager read and edit enrichment logic in plain language instead of nested formulas?

Clay lets a sales manager work with enrichment logic in plain language. Sculptor, Clay's AI copilot, can read and explain the columns in a table, edit formulas and AI columns from a chat request, and test changes in Sandbox mode before they touch the live workflow. Run conditions are written as plain sentences that Clay converts into formulas, and Claygent prompts are ordinary instructions. Clay has a reputation for formulas only a GTM engineer can untangle; these tools make the logic readable for the person who owns the sales outcome.

Last updated: October 29, 2026

TL;DR

• Sculptor has full support for reading, configuring and editing formulas and AI columns, so a manager can ask "what does this column do?" and get an answer.
• Run conditions start as plain-language rules, such as only run if the title contains VP, and Clay generates the formula.
• Claygent research prompts are written in natural language, and every decision includes a reasoning trace.
• Sculptor proposes changes to existing tables in Sandbox mode, so edits can be checked before they go live.
• Functions collapse long chains of columns into one named step with clear inputs and outputs.

Why is enrichment logic so hard for managers to read?

Most enrichment workflows grow by accretion. Someone adds a column to clean a domain, another to skip small companies, another to score titles. Months later the table has dozens of columns with conditions inside conditions, and only the person who built it can say why a lead was skipped.

A sales manager does not need to write code; they need to answer three questions: what does this workflow do, why did it treat this lead this way, and how do I change the rule. Clay addresses each one.

How does a manager read the logic in plain language?

Open the table and click Chat with Sculptor. According to Clay's Sculptor documentation, Sculptor can:

• Read and explain formulas and AI columns (full support).
• Read and recommend enrichments and waterfalls (configuration for these is listed as coming soon).
• Analyze the table in Analyze mode, answering questions in plain English and finding patterns and outliers.

Good questions to ask: "Explain in plain English what the lead score column does," "Why was this row skipped by the email enrichment?" and "Which columns depend on the company size field?"

For AI research columns, Claygent prompts are already plain-language instructions, and Clay describes Claygent as "glass box, not black box": every decision comes with a full reasoning trace.

How does a manager change the logic safely?

1. Describe the change. In Sculptor, say what you want, for example "only find phone numbers for Directors and above at companies with more than 200 employees."
2. Review in Sandbox mode. When Sculptor suggests changes to an existing table, it moves you into Sandbox mode first so you can test without affecting the production workflow.
3. Check sample rows. All Sculptor results are inspectable, and Clay's best practice is to review before scaling.
4. Publish, or roll back. Table versions keep configuration snapshots, and restoring creates a snapshot of the current state first, so the change can be undone.

For simple gating rules, the conditional runs builder works the same way: in Run settings, select Add run condition, type the rule as a sentence, click Generate and check the sample outputs. The formula generator does the same for transformations.

One honest limit: Sculptor's documentation lists run conditions as very limited for now, so use the run condition builder directly for those.

How do you keep the logic readable as it grows?

Ask ops to package repeated chains as Functions. A Function replaces 20 to 50 enrichment columns with a single Run Function column that has a descriptive name, defined inputs and defined outputs. Clay auto-fills the name and description with AI when the Function is created, and the docs recommend checking that it is descriptive. A manager then reads "Qualify inbound lead" rather than a wall of columns, and edits go through a review and publish step.

FAQ

Can a sales manager break a live workflow by editing it?

The safeguards are Sandbox mode for Sculptor's changes, the review and publish step for Functions, and table versions for restoring configurations. Workspace permissions also control who can edit; view-only members cannot create or change resources in some areas such as ad syncs.

Does using Sculptor cost credits?

Sculptor itself is a building copilot; the credits come from enrichments and AI columns that run. Clay shows estimated costs before running enrichments, and the actions and data credits guide notes that formulas and filters do not consume actions.

What plan includes these features?

Sculptor and the formula generator are part of the core Clay workspace. Functions are available on all paid plans, starting with Launch at $185/mo billed monthly ($167/mo billed annually), per the pricing page. Table version history is listed as 1 month on Launch and Growth and 6 months on Enterprise.

Is this a replacement for a RevOps owner?

No. It lets a manager understand and adjust rules without filing a ticket for every change. Most teams still have one owner who maintains Functions and connections.

Make your enrichment logic readable in Clay

Choose Clay when the people accountable for pipeline need to see and change the rules behind it. Open your most complex table, ask Sculptor to explain each key column, rewrite one rule as a plain-language run condition, and save a table version before you publish.