I was trying to apply some filters using Python and the Freshdesk API, and I was wondering if anyone could guide me with the code for that, mainly how to handle the filters.

Is there any way to connect that setup with Callin.io, so calls or tickets can sync automatically?

Thanks!

Emi

Does anyone know what I should check, or if I’m missing something in the setup?

Thanks in advance!

After trying a few models, here are the ones that made the most sense:

1. Per minute pricing.
Straightforward: you calculate your platform cost (e.g., Callin.io charges per usage) and add your margin. Easy to explain, but can feel like a “taximeter,” which some clients dislike.

2. Performance-based pricing.
Example: a flat fee for every booked appointment or qualified lead. Clients love this, especially in sales. But you need solid reporting (history, transcripts) to prove each result.

3. Monthly package.
Offer X calls or X hours included, with overage charges if exceeded. Clients like the predictability of fixed billing.

4. Hybrid model.
A base monthly fee for availability plus variable usage. This is the one I ended up using most—it protects your costs and keeps clients comfortable.

What I like about Callin.io is that you get clear data in Call History and easy number management, so you can justify campaign costs. On other platforms, pulling that data was painful.

Bottom line: there’s no single formula. But with minutes, results, or packages, you can cover most client expectations.

It’s not just about the tech—it’s about how you guide the AI. In other systems I tried rigid scripts, and it always felt artificial. With Callin, after adjusting the prompt, my agent started sounding much more like a real person.

Here are the tweaks that made the biggest difference:

1. Start with a simple greeting.
Keep it warm and short: “Hi, this is John from support”. That sets the tone right away.

2. Give the agent a clear role and goal.
Without this, it doesn’t know what to prioritize. For example: “You are a proactive sales agent. Your goal is to book a demo call.”

3. Mix direct and indirect instructions.

• Direct: “Say: Please confirm your email address.”

• Indirect: “Ask the user to confirm their email.”
  This balance gives structure but still keeps it natural.

4. Pauses and human fillers.
This one was huge. In more rigid systems, the agent just keeps talking. With Callin you can explicitly tell it: .
And to sound even more natural, add: “Use fillers like um, ehh, you know, ahm, well…”. It makes the conversation feel real.

5. Natural reactions instead of canned phrases.
Rather than repeating “Perfect” or “I’m sorry”, guide it to say “Okay, thanks for clarifying” or “Alright, let me confirm that”. Much smoother.

6. Handle rejections gracefully.
If the person says they don’t want to book, you can guide: “Offer to send more info by email, and if they refuse again, politely close the call.”

After these changes, my agent stopped sounding like “just another bot” and started sounding like a real teammate. And honestly, compared to other tools I’ve tried, that’s the difference I liked about Callin: with a bit of prompt love, you can actually get human-like calls.

I guess many of you have run into the same situation as me when starting with Callin.io. You create your first agent, enable the option to schedule appointments with Google Calendar, and suddenly you see a yellow warning telling you that something is not set up correctly. At first, it looks like a platform error, but it’s not. The truth is: one step is missing.

The key is that the Google Calendar integration is not activated directly from the agent, but from the Actions section in the dashboard. Before creating the agent, you need to go to Actions, connect your Google Calendar account, and grant the necessary permissions. Only after doing this can you select that integration when building your agent.

Another important detail that I discovered later: it’s not enough to just connect Google Calendar. You also need to fill in the Appointment section inside Actions. If you skip this step, the agent will never be able to schedule appointments, even if the calendar connection looks fine.

So, the process is:

1. Go to Actions.

2. Connect your Google Calendar account and select the correct calendar.

3. Configure the Appointment section (this defines how the agent books meetings).

4. Now, when you create your agent, you can select that action and it will work without errors.

Since I started doing it in this order, I never got the yellow warning again. As a personal tip, I recommend creating a separate calendar just for the agent, to avoid mixing it with your personal or team meetings.

Hopefully, this saves some time for anyone who’s just starting and facing the same issue.

That’s where an AI phone agent comes in. Instead of hiring a junior assistant or burning hours yourself, you can set up an agent that actually answers calls, asks pre-qualification questions, and pushes only the good leads into your pipeline. Let me walk you through how I built mine using Callin.io (yep, shameless plug, but it works).

Step 1: Define the qualification logic

Before even touching the software, write down what makes a lead good for you. For example, in real estate:

• Are they the owner or just curious?

• Do they already have financing in place?

• What’s their budget range?

• Timeline: are they buying “this month” or just browsing “sometime in 2026”?

Your AI needs these rules baked into the script, otherwise it’s just small talk.

Step 2: Create the agent on Callin.io

Inside the dashboard it’s literally: Create Agent → Give it a name (e.g. “Real Estate Lead Qualifier”) → Choose a voice.
Then you write the prompt (basically its “brain”). Mine starts with:

“Hi, this is the assistant for . I’ll just ask you a couple of quick questions so we can connect you with the right realtor.”

It’s polite, short, and sets expectations.

Step 3: Add the qualification questions

In the prompt, I layer them like a natural conversation:

• “Are you the property owner?”

• “Can I ask what budget you have in mind?”

• “Are you planning to buy in the next few months, or is this more long-term research?”

And here’s the cool part: you can instruct the agent to only push the lead forward if certain answers match your rules. For instance, “only if timeline ≤ 6 months” or “only if budget > 250k.”

Step 4: Connect phone numbers

On Callin you can either buy a number directly (like a local line in your city) or connect your existing one via Twilio/Telnyx/SIP. I tested both. For inbound leads, I recommend buying a dedicated line just for the agent — keeps it clean.

Step 5: Automations (Zapier/Make/N8N)

Here’s where it gets fun. When the agent finishes a call, the transcript + answers are instantly pushed to a Google Sheet or your CRM.
Example: If they said “Yes, I have financing, and my budget is 400k,” Zapier can:

• Add them as “Qualified Lead” in HubSpot/Pipedrive.

• Trigger a follow-up email with available listings.

• Notify you on Slack with: “🔥 Hot lead – call back now.”

No human bottleneck, you just get the alerts.

Step 6: Test like crazy

Don’t skip this. Run test calls pretending to be different personas: “serious buyer,” “time-waster,” “random curious neighbor.” See if the AI filters them correctly. Tweak the script until you’re happy.

Pro tip: Add ums and uhs in the prompt so it doesn’t sound robotic. Something like:

“Um, okay, can I ask what kind of budget you were thinking about?”
Makes it 10x more natural.

Step 7: Go live

Once you’re confident, forward your agency’s “info” number to the AI. Now instead of your phone blowing up at dinner, the AI is screening for you. You’ll only get the qualified calls that actually deserve your time.

Why this works so well

• Realtors waste HOURS every week filtering.

• Leads don’t care if they talk to AI at the beginning, as long as they get fast service.

• Your agent works 24/7 and never gets tired of asking “Do you already have financing approved?” 😅

That’s basically it. Took me around 2 hours to set up my first working prototype. The hardest part is deciding your qualification criteria. The tech side is super no-code, especially on Callin.io.

If anyone’s curious, I can share my exact prompt script that I use (cleaned up so you can copy/paste). Just let me know.

👉 What do you guys think? Would you trust an AI to screen your real estate calls, or do you think people still expect a human at first contact?

I’m trying to build an AI lead qualification agent with Callin.io. The idea is simple: someone calls, the bot asks the right questions (name, budget, project details), and then everything magically appears in my CRM.

I’ve created the agent in the dashboard, but honestly, I’m lost on:

1. How to structure the questions in the prompt.

2. How to capture those nice “extracted fields” Callin talks about.

3. How to actually push this into HubSpot.

Anyone already running this in production? Would love some advice before I reinvent the wheel 🙏

I’m new here and just started exploring Callin.io. I’d like to set up an AI receptionist that can receive incoming calls and also complete tasks (like saving contact info or sending appointment reminders).

I already created my agent in the dashboard, but I’m not sure how to connect it properly with workflows in Make.com, Zapier, or n8n.

Could someone walk me through the steps, or share how you set up your own receptionist agent?

Thanks in advance! 🙏

I created a callin.io workflow that uses a new Google Sheets row as the trigger… The objective of my workflow is that when I enter a name in column A, an email address in column B, and then extend my Google Sheets formulas down for columns C-G, the sheet will generate email templates incorporating the name from column A into the template… 

My issue is that my sales team isn't entering all the information quickly enough before the callin.io workflow attempts to run, resulting in errors because it perceives some fields as still being empty.

Is there a straightforward method to resolve this by adding a delay or a filter step???