Kiln — Waterfall Enrichment & AI Research Engine
Know what enrichment costs before you run it. Billed only for verified results.
Simulate Your Outbound List Economics
Yields 2,250 verified email results.
The Predatory Economics of Legacy Enrichment
Go-to-market and outbound teams waste fortunes on dirty contact data. Legacy enrichment platforms charge 4 to 8 credits ($0.20 to $0.40) per verified email, charge upfront even when lookups fail or time out, and lock essential features like CRM sync and HTTP APIs behind $495+/month enterprise paywalls. Furthermore, waterfalls lack real-time price predictability, leading to surprise monthly billing spikes.
Transparent Waterfalling & Zero-Waste Billing
Built Kiln from the ground up as a high-throughput, cost-transparent waterfall enrichment and AI research engine. Kiln chains 12+ provider adapters, forecasts the exact dollar cost and match rate before credits move, and enforces deterministic verified-only billing: misses, provider errors, and cached repeats are 100% free. Contact enrichment drops to ~4¢ per verified email (an 81% cost reduction) with full CRM sync and HTTP API included on the entry $49 tier.
Core Innovations
Engineered for scale, speed, and safety.
Deterministic Verified-Only Billing
The core economic engine operates on `billed = hit && verified && !cached && !sample`. If a provider returns no match, encounters a rate limit, or fails MX/SMTP verification, the customer pays zero.
Pre-Run Cost & Match Forecaster
Before a single credit is deducted, Kiln's mathematical estimator predicts total cost, expected match percentage, and provider hit distribution based on historical confidence curves.
Schema-Enforced AI Research Columns
Enrich beyond emails by asking questions of every row in plain English (e.g. 'Does this company sell to developers?'). Structured JSON schemas guarantee clean inputs with zero hallucinations or malformed rows.
Bring Your Own Keys (BYO Keys)
Plug in private API keys for Hunter, Prospeo, Findymail, or Dropcontact on any plan. Users pay the data vendor directly at cost while Kiln charges a fractional orchestration fee on verified hits.
Un-gated Native CRM & Webhook Sync
Direct bidirectional syncing into HubSpot, Salesforce, Instantly, Smartlead, and Slack available on the $49/mo Starter tier rather than gated 4 tiers up.
Server-Side Hard Spend Caps & GDPR by Design
Active server-side spend caps prevent accidental overages. Suppression lists halt provider network requests before invocation, eliminating compliance risks and wasteful spend.
System Architecture
How the system works under the hood.
1. Waterfall Pipeline & Provider Orchestration
Chains up to 12 provider adapters (Hunter, Apollo, Clearbit, Prospeo, Findymail, Icypeas, Dropcontact, etc.) in user-specified precedence. Executes cascading requests asynchronously and short-circuits at the first verified positive hit, minimizing API overhead and token consumption.
// Provenance & short-circuiting waterfall step runner
async function executeWaterfallStep(contact: ContactInput, chain: ProviderAdapter[]): Promise<EnrichmentResult> {
const cached = await cacheStore.get(contact.domain, contact.fullName);
if (cached) return { ...cached, billed: false, reason: "cache_hit" };
for (const provider of chain) {
if (await suppressionList.matches(contact)) break;
const res = await provider.lookup(contact);
if (res.status === "HIT" && res.verified) {
await logProvenance({ provider: provider.id, cost: provider.unitCost, confidence: res.score });
return { email: res.email, billed: true, provider: provider.id };
}
}
return { status: "NO_MATCH", billed: false }; // Misses are 100% free
}2. The Deterministic Billing Engine
Guarantees that customers only pay when tangible value is returned. The ledger operates with atomic transactions linked to provenance logs, ensuring transparent accounting.
// Deterministic billing formula enforced at transaction boundary
const shouldBill = hit && verified && !cached && !sample && !suppressed;
if (shouldBill) {
await ledger.debitCredits({
workspaceId,
amount: providerRate,
idempotencyKey: `enrich_${rowId}_${providerId}`
});
}3. Schema-Validated AI Agent Execution
Agent research columns execute parallel inference passes with structured outputs. Prompts define strict JSON schemas (such as { fits_icp: boolean, justification: string, score: number }), ensuring malformed LLM outputs never corrupt the spreadsheet.
// Schema enforcement for row-level AI research
const agentSchema = z.object({
sells_to_devs: z.boolean(),
evidence: z.string().max(180),
confidence: z.number().min(0).max(1)
});
const response = await aiClient.generateStructured({
model: "gemini-2.5-flash",
schema: agentSchema,
prompt: `Analyze website content for ${domain}...`
});Direct Comparison
Where Kiln is structurally different.
| Capability | Kiln | Clay | Spreadsheet + VA |
|---|---|---|---|
| Billing on Misses / Errors / Timeouts | Free ($0.00) | Free (since Mar 2026) | Manual VA hours billed |
| Cost of Verified Work Email | ~4¢ (less with BYO key) | 20¢–40¢ (4–8 credits) | $0.50–$1.50 per record |
| Pre-Run Match Rate & Cost Forecast | Full forecast before run | Per-cell estimate only | None / Post-hoc invoices |
| CRM Sync & HTTP API Access | Included from $49/mo | Gated from $495/mo | Manual CSV exports |
| Bring Your Own Provider Keys (BYOK) | Every tier (pay at cost) | Higher tiers only | N/A |
| Server-Side Hard Spend Caps | On by default at allowance | Opt-in budget alerts | None |
| 5,000 Row List Turnaround | Minutes (<180s) | Minutes | 3–5 business days |
Technology Stack
Built with precision tooling.
Need a production system built with this level of rigor?
I partner with founders and enterprise teams to architect high-throughput SaaS engines, reliable AI pipelines, and rock-solid web infrastructure.
Book a 30-min Architecture Review