The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
AI Customer Support Agent for Web Form, Gmail & WhatsApp
An AI "digital employee" that answers customer support questions across a web form, Gmail and WhatsApp, and hands off to a human team when it should. Load-tested with 200+ messages and no lost messages.
I built an AI customer support agent as a demo for a fictional SaaS company called Flowmint. Customers can write through a web form, Gmail or WhatsApp. The agent finds the answer in a knowledge base, replies on the same channel, and passes the conversation to a human team (Billing, Sales, Support) for refunds, special pricing, or when a customer asks for a person.
One agent for three channels, with replies formatted for each (a clean HTML email, a short WhatsApp message, a private web tracking page)
Safety rules written in code: it can only quote approved prices, and refunds always go to a human
Remembers returning customers on email and WhatsApp
Private ticket tracking page with a secret tracking code
If the AI service is busy, it switches to backup AI models, and to a knowledge-base answer if all are limited
I load-tested the agent for 24 hours with 200+ messages across all three channels, using simulated customers plus real emails and WhatsApp messages I sent myself.
212 messages in, 212 replies out, none lost
Only 3.3% needed a human
In a separate Kubernetes test, I stopped the worker three times while messages kept arriving, and every message was still answered exactly once
Areas I'm improving: faster replies with a paid AI plan, and spam protection for the web form.
Tech stack:
Python, FastAPI, OpenAI Agents SDK, Kafka, PostgreSQL, Next.js, Docker, Kubernetes, Gmail API, WhatsApp Cloud API.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
Meet Moe.
Moe is the mascot and AI guide we designed for our studio's website. We started with the character: a soft, friendly cube, modelled, lit and rendered in Blender. All six expressions (happy, wink, wow, thinking, chill, curious) are posed from the same 3D model, so he stays consistent across the site and our socials.
On the site, Moe is rebuilt in code. He waits in the corner, blinks, follows your cursor with his eyes, and dozes off when ignored, then opens into a chat that helps founders work out what they need: which package fits, what's included, the timeline and how soon work can start. His answers come straight from the site's CMS, so they stay current when pricing changes, and he hands off to the team when a question needs a person.
Need a mascot for your brand or an AI assistant on your website? Send us a message here; we'd love to build yours.
That 3D mascot branding combined with an AI-powered product studio concept is brilliant, Ibrahim! The clean UI and character expressions give the whole agency presence a ton of personality.
Everyone is shipping AI agents right now. Almost no one is testing them — and that is the whole problem.
A demo agent and a production agent are different animals. The demo works because you walked it down the happy path. Production breaks on the paths you never imagined.
A real example from my own work: a support assistant correctly handed a conversation to a human when someone wrote "I want to talk to a person" in Spanish — but not in English, because one phrase was missing from a rule. Every automated test passed. It only surfaced when I ran real conversations through it.
That is the gap the "agents will replace everything" wave keeps skipping. The hard part was never making an agent act. It is making it:
• say "I don't know" instead of guessing,
• hand control back to a human at the right moment,
• and prove, with tests, that it still behaves when the input gets strange.
Autonomy is cheap. Trustworthy autonomy is the real engineering.
Before you let an AI agent near real users or real data — what is the first thing you check?
Most inbound bottlenecks aren't caused by low lead volume—they happen during triage.
When inquiries roll into a Google Sheet or CRM, operations teams often waste 10–15 minutes per submission reading through messy paragraphs, deciphering budget mismatch, and manually deciding who gets an immediate call vs. a standard rate sheet.
Instead of adding a bulky CRM add-on, I documented a lightweight 3-step automation pipeline built inside Make:
Intake & Extraction: An active listener triggers immediately upon a new submission, extracting core variables: client scope, target deliverables, timeline, and stated budget.
Context-Aware Evaluation (Gemini 2.5 Flash): The payload passes through an AI evaluation node with strict guardrails:
Evaluates budget against minimum agency criteria.
Assigns a triage score: Hot (immediate priority), Warm (nurture), or Cold (disqualified/template reply).
Generates a strict 15-word operational summary—no long paragraphs.
Determines an automated next step (e.g., routing to an account lead or flagging for asynchronous follow-up).
Data Return & Clean Pipeline: Appends the structured classification directly back to the master operational tracker, keeping the sheet filterable and clean.
The outcome: The sales and operations team can filter the entire board by priority status in two clicks, cutting first-touch response time to minutes without manually reading every single submission.
I’ve documented the full scenario map, execution logic, and data schemas below.
Curious how other agency operators and automation leads here handle their lead enrichment layer—are you doing AI scoring at the form level, or later in the CRM pipeline?:
For the budget gate, I'd keep an unstated budget separate from one explicitly below your minimum. Does the Make route ask for clarification when the amount or currency is missing, so a potentially suitable lead doesn't fall into Cold by default?