Your AI sales agent should not invent discounts.
Chatbots and AI sales reps quote prices to customers every day. One API call before each price keeps them inside your rules: floors stay secret, deep discounts go to a person, and every decision lands in a tamper-evident log.
POST /v1/guard/check
X-Guard-Key: gk_…
{"sku": "SENSOR-V2", "qty": 120,
"unit_price": 250, "customer": "Gamma Plant GmbH"}
→ {"decision": "deny",
"suggested": 289.9,
"reason": "Counter-offer"}Three answers, decided by your rules
The decision comes from your price list and limits, not from a language model. The same request always gets the same answer.
The price is within the discount your agent may give on its own. Say it to the customer.
Deeper discount or a large deal. Your sales lead gets a message and approves or rejects in the console with one click; your agent polls the result.
Below your floor. The answer includes a price your agent can offer instead, never the floor itself.
Ten lines in your agent
Call it as a tool from any agent framework: OpenAI, Anthropic, LangChain or your own code.
import httpx
def check_price(sku, qty, unit_price, customer=None):
r = httpx.post("https://quotegateai.com/v1/guard/check",
headers={"X-Guard-Key": GUARD_KEY},
json={"sku": sku, "qty": qty, "unit_price": unit_price, "customer": customer})
return r.json() # decision: allow | needs_approval | deny
# needs_approval → poll GET /v1/guard/checks/{id} until status is approved or rejectedPricing
We load your price list and limits in one call. Extra checks are $5 per 10,000.
Price guardrails for your own AI sales agent
one API call before it names a price returns allow, needs approval or deny, and logs every decision. Includes 10,000 checks a month$299/moTalk to usGet an API key
Tell us which agent quotes your prices. We set up your rules and send a test key within one business day.