What an AI lead-qualification agent actually costs

A customer support bot runs when someone chooses to open the chat widget. A lead-qualification agent runs on every inbound form submission. That distinction changes the cost math completely.

What a lead-qual agent actually does

When a new lead comes in — contact form, landing page, ad click — the agent pulls the submission, makes an LLM call to score and categorize it (product fit, urgency, company size signals, whatever criteria matter to your sales team), then fires a webhook to your CRM to tag and route the lead. Optional: trigger an SMS follow-up via Twilio if the score crosses a threshold.

Three cost lines run every single time a lead comes in.

Cost line 1: The LLM scoring call

A well-scoped lead-qual prompt — system instructions plus form fields plus scoring rubric plus structured output — runs around 1,500–2,500 tokens per call. At Claude Haiku rates ($0.80/M input, $4/M output), that's roughly $0.003–$0.005 per lead. At Sonnet ($3/M input, $15/M output), it's $0.006–$0.012 per lead.

At 500 leads/month: Haiku costs about $1.50–$2.50/month in LLM spend. Sonnet costs $3–$6/month.

At 5,000 leads/month: Haiku is $15–$25/month. Sonnet is $30–$60/month.

Those numbers are manageable — until you add the next two lines.

Cost line 2: CRM webhooks and third-party data enrichment

The webhook call to your CRM (HubSpot, Salesforce, Pipedrive, whatever) is usually free or included in your plan. But enrichment — looking up the company in Clearbit or Apollo to fill in firmographic data before the LLM scores the lead — can cost $0.01–$0.05 per lookup depending on provider and plan tier.

At 5,000 leads/month with enrichment at $0.02/lead: add $100/month to the stack.

If you're using a reverse IP lookup or real-time LinkedIn enrichment, that number goes higher.

Cost line 3: SMS follow-up (Twilio)

Twilio SMS in the US runs about $0.0079 per message outbound. If every qualified lead (let's say 30% of submissions) gets an automated follow-up:

• •500 leads/month × 30% × $0.0079 = $1.19/month
• •5,000 leads/month × 30% × $0.0079 = $11.85/month

Not huge, but not zero. At 25,000 leads/month — common for e-commerce or high-volume B2C — you're at $59/month just in SMS.

Why volume sensitivity matters more here than in support bots

A support bot scales with how often customers have problems. Lead volume scales with how much you spend on ads and how well your site converts. A good paid media campaign can double your inbound leads in a week. If your lead-qual agent is running on a fixed API key with no spend cap, a successful campaign is also a surprise bill.

We always set spend limits and alert thresholds on lead-qual deployments. We also model three scenarios during discovery: your current volume, 3x volume, and 10x volume. If the cost at 10x is a problem, we know before we build, not after.

The full stack at two volumes

| Component | 500 leads/mo | 5,000 leads/mo | |---|---|---| | LLM scoring (Haiku) | $2/mo | $20/mo | | CRM webhooks | $0 | $0 | | Enrichment (optional) | $10/mo | $100/mo | | SMS follow-up (30%) | $1.19/mo | $11.85/mo | | Total (no enrichment) | ~$3/mo | ~$32/mo | | Total (with enrichment) | ~$13/mo | ~$132/mo |

The build to set this up — prompt engineering, CRM integration, scoring rubric, testing — is typically $2,500–$6,000 depending on how custom the scoring logic is and how many CRM objects need updating.

Care tier covers ongoing prompt tuning and monitoring. API spend is always pass-through.

You can model your specific volume at our cost estimator. If your numbers look different from these, that's worth understanding before you sign anything.