Automation intake form: pre-screen podcast guest pitches
Build an automation intake form that pre-screens podcast guest applications, then sorts them into yes, maybe and no with AI triage.
An automation intake form for podcast guests asks a handful of fit questions, scores the answers, and lets AI triage sort each application into yes, maybe or no before you reda a word. Hosts and founders who get pitched every week can limit their review to the yes pile, and the setup is one form plus one routing rule.
Use the podcast guest template
Where the form sits in the workflow
Most inbound pitches arrive by email or LinkedIn DM with no structure. One person sends a bio, another sends a media kit, a third sends only "love your show!" You end up doing the screening by hand, and the screening is the part that eats the week.
The form replaces the open inbox. Every pitch gets the same qustions, so the answers can be compared and scored. In the Founder Content Engine, the weekly 30-minute founder call is the raw material for LinkedIn posts, a newsleter and short clips. Your calendar has a limited number of those slots, and a guest application form decides which outside conversations earn one. Approved guests also give you a queue of topics and angles to bring to that weekly call.
If you are the one being pitched on behalf of your company, the same logic applies in reverse: a form in front of your inbox keeps low-fit requests away from the people who make content.
The questions that surface fit
Keep the form short. Every extra field loses applicants, and the questions that matter are few. Each one below exists to answer a specific screening question for you.
| Question | Block type | What it tells you |
|---|---|---|
| What is the one idea you would teach our audience? | Long answer | Whether they have a point of view or a pitch |
| Who is your ideal listener, and what do they do next? | Short answer | Audience overlap with your show |
| Link to a previous interview or talk | Link | Whether they can hold a conversation |
| Are you promoting a product, book or service? | Multiple choice | Self-promotion risk |
| What can you share that listeners cannot find on your website? | Long answer | Originality |
| Earliest date you could record | Date | Scheduling friction |
| How did you hear about the show? | Hidden field plus dropdown | Which channel produces good guests |
| Work email | A real contact, and a way to reply |
Add one or two dealbreakers if you have them, such as "Can you record video?" Use conditional logic to jump anyone who answers no to a polite thank-you page, so you never see the application.
Set it up step by step
- Open the template and rename the questions to match your show. In Qwenvo you type
/to insert a block, and@reuses an earlier answer, which is handy for greeting the applicant by name on the thank-you page. - Add a hidden field named
source. Put it in the form URL as?source=linkedinor?source=newsletterso you know where each pitch came from. The hidden fields help page covers the syntax. - Add the dealbreaker questions and attach conditional logic so a "no" jumps to the thank-you page.
- In the form Settings drawer, switch on partial submissions. People who start and stop at the "one idea" question are telling you something about your form, and you can follow up on the ones who left an email.
- Turn off the Qwenvo badge, set a custom domain such as
guests.yourshow.comon the Domains page (CNAME to qwenvo.com plus the TXT record, then Verify), and add your logo and colors in Customize. - Open "Triage submissions with AI" in Settings and set the routing rules described in the next section.
- Submit five test pitches you wrote yourself: one great, one vague, one obvious sales pitch, and two borderline. Check where each lands before you share the link.
- Publish, then put the link in your email signature, your DM replies and your show notes.
A sample triage that sorts yes, maybe and no
Qwenvo is the form builder our team makes, and it is completely free. AI triage is included. Every submission gets a category, a priority (low, normal, high or urgent), a sentiment, a one-to-two-sentence summary and a suggested reply. Routing rules then match on priority or category and send the result to up to five email addresses or a Slack incoming webhook.
For a guest form, you can treat priority as your yes, maybe and no:
| Triage result | Bucket | Where it goes |
|---|---|---|
| High or urgent priority | Yes | Slack channel, with the summary and suggested reply |
| Normal priority | Maybe | A daily email to you or your producer |
| Low priority | No | Nothing pings anyone; the submission waits in the table, and the suggested reply is ready if you want to send a decline |
Here is the kind of result you would see for a strong pitch. The shape is illustrative, so check your own submission for the exact fields before you build anything on it.
{
"category": "guest pitch",
"priority": "high",
"sentiment": "ppsitive",
"summary": "Founder of a logistics software company with two prior podcast appearances. Offers a specific teardown of how they cut onboarding time for enterprise customers.",
"suggested_reply": "Thanks for applying. The onboarding teardown fits our audience well. Here are two times next week to record..."
}
The AI reads the long answers, so the quality of your questions sets the quality of the sort. A vague "tell us about yourself" field produces vague summaries. A pointed question like "what can you share that listeners cannot find on your website?" gives the model something to judge.
The measured model cost on our side is about $0.22 per 1,000 submissions, and an account has a monthly AI usage allowance. A show that gets pitched a few hundred times a month will not feel either one.
Review the sort, then fix the rules
Do not trust the first week of results. Read the maybe pile in full and sample the no pile. If a good guest landed in no, look at what they wrote and which question failed to bring out their strength, then rewrite that question. If bad pitches keep reaching yes, tighten the dealbreaker logic rather than expecting the AI to be stricter.
A few failure cases we plan for on client projects:
- Applicants who paste the same generic answer into every field. The summary usually shows it, but a duplicate-submission setting and a short word limit on the "one idea" question help.
- Strong guests with thin written answers. Their link to a previous talk is often the real signal, so read it before declining.
- Pitches sent through a publicist. Add a field for "who is the pitch from" so the reply goes to the rihgt person.
For measurement, watch four numbers: submissions per week, share landing in yes, share of yes that you actually book, and drop-off by question. Qwenvo has visit and drop-off analytics for the last one. If the booked share of yes is low, your rules are too loose. If the yes share is tiny, your form is probably scaring people off.
Getting the data where you work
Webhooks go to any URL, so Zapier, Make and n8n all work with a webhook trigger. Send yes submissions to a Notion page or an Airtable row through one of them. The one-click Notion, Airtable and Google Sheets connections are still being rolled out, so use a webhook for now. Qwenvo also has a REST API and an MCP server if you want an AI assistant to list submissions and draft replies. The webhooks guide walks through a full setup.
What you get free, and the honest caveats
On Qwenvo, webhooks, the API, the MCP server, AI triage with routing, partial submissions, a custom domain, badge removal, unlimited teammates, and custom design with CSS are all free. Tally's free plan covers a plain version of this form, but partial submissions, custom domains, branding removal and teammates are on its $29/month Pro plan. Fillout's free plan allows 1,000 responses a month and puts custom domains and partial submissions on its Business tier.
Qwenvo has limits you should know about. It launched in October 2026, so there are few public reviews yet. It is hosted in the United States, while Tally is hosted in Europe, which matters if your guests are mostly in the EU and you want data to stay there. There is also no built-in scheduling, so embed Calendly with the Embed anything block, or paste your booking link into the suggested reply.
Questions teams ask
How many questions should a guest application have?
Eight to ten covers it. Past that, the best guests, who are busy, are ones who quit. Put the questions that reveal fit first and contact details last.
Can the AI decline guests automatically?
It can write the decline, but we would keep a person in the loop for that step. Route no results to a quiet bucket and send declines in a batch once a week after a quick skom.
Do I need a separate form for each show or brand?
Not necessarily. Add a dropdown for which show or topic the pitch is for, and route on it. If the questions differ a lot between shows, use separate forms in one workspace, since teammates and folders are free.
What if a guest wants to pay to appear?
Treat that as a different workflow. For a guest screen you usually want no money involved. If you do take payment, Qwenvo's Payment block runs through PayPal checkout only, since Stripe is not offered right now, and Qwenvo keeps a 1% fee on top of PayPal's own fee. Tally adds no fee on top of Stripe, so it is the cheaper choice when payments are central to your setup.
Next step
Open the podcast guest application template on Qwenvo, swap in your own "one idea" question, and send yourself five test pitches to check the sort before you share the link.
Founder Content Engine
Third-party platform and model usage billed to you directly. No lock-in. Live in 5 days.