How we build a prospecting tool with the reps who use it

A good prospecting tool is built with the sales reps who use it every day, not in their place. At Flunter, we start from a need seen in the field, measure how often it comes up and how much time it costs, then ship a first version quickly. One rule never changes: the tool saves time around the call, but the conversation stays human.

Flunter is a parallel dialer: it calls up to 5 prospects at once and only passes the rep the ones who answer. Running several calls in parallel sharply increases the number of conversations per hour. But it is not the only lever.

What I see in the teams we work with is that time also slips away in small frictions: finding a note before calling back, listening to a call again, looking for a contact, retyping something into the CRM. Each one seems minor. Added up over a day, they cost calls.

That is why our comparison page with Orum says this, and I fully stand behind it:

"We don't aim to have the most features. We aim to build the ones your sales reps will use every single day." (Flunter vs Orum)

In practice, before adding anything, we ask one question: will a rep use it at least once per calling session? If not, the idea waits.

Needs reach us through very simple channels, and that is on purpose.

• Questions at the end of my LinkedIn posts. When I present something new, I sometimes end with a question. On May 4, for example: "How do you handle your calls after a session?". The answers tell us what really gets in the way.

• Direct conversations with teams. In-app chat, direct messages and demos are where we hear requests in the field's own words: "I want to quickly find a contact's notes before calling them back", "I want to dial a number directly".

• Public feedback from users. Some share their results on LinkedIn, and I reshare them. They show us what matters to them.

Two public examples I often mention. Priscilla De La Rosa prospects for insurance brokers. Before, she described "voicemail after voicemail, a huge waste of time" (her post was in French, as was Jeremy's). After her first day on Flunter, in December 2025, she announced 4 qualified meetings for a broker in the Tarn area and 2 for a broker in Bordeaux. Jeremy L., founder of JL3P with 15 years of B2B phoning behind him, posted in January 2026: "9 meetings booked in 2 hours of cold calling, thanks to Flunter".

These are their results, not an average. But they remind us what reps expect from a tool: less time lost on voicemail, more real conversations. Our voicemail detection runs on an in-house engine, tuned continuously on real calls, that decides at pickup.

Here is what came out between late April and late May 2026, with the need behind each one.

The API is a good example. Teams that prospect a lot never work in a single tool: they have a CRM, homemade dashboards, internal workflows, and I see more and more tools built on Airtable. When Flunter becomes the calling engine, it has to fit cleanly into that setup, on top of the native CRM integrations. We first opened it to beta testers; API access is now included in every plan.

For the contact record, the need was simple: before calling someone back, you want to know what was said. The AI summary and recording with transcription already existed, but you had to dig them out call by call. Bringing them together on the record solved it.

The call companion is a first step. The idea is to make it a point of support that gradually helps reps prospect with more context and less friction.

We get more requests than we can ship. To sort them, we use a deliberately simple rule: how many times the need comes up in a day of prospecting, multiplied by the time it costs each time.

Here is a worked example, for illustration only. If a rep spends 30 seconds finding a contact's context and does it 20 times per session, that is 10 minutes lost every session. For a team of five people doing one session a day, that comes to about 4 hours a week. On the other hand, a very detailed report checked once a month can be useful, but it does not create a single extra conversation.

This rule has its limits. A rare but blocking need, like an import that fails, jumps ahead of everything else. And some requests come from a single customer but reveal a problem others live with without saying so.

Saying no is part of the job. Here is what we openly refuse to build today.

• An AI voice that talks instead of the rep. The idea comes up often in the market. For us, it's a no: Flunter is not an AI that talks to prospects. As soon as a prospect answers, the rep does the talking. The tool works around the call (voicemail detection, transcription, AI summary), not in place of the rep.

• Automated calls without human control. Flunter does not allow automating calls without human control. The parallel dialer dials several numbers, but every conversation is led by a person.

• Prerecorded voicemail drops. Flunter does not leave prerecorded voicemail messages. A prospect who gets an automated message does not have the same experience as a conversation with a real rep.

• Complexity for its own sake. A good prospecting tool should not turn into a complicated machine reserved for large teams. Every new screen has to be justified by daily use.

These refusals have a cost: sometimes we miss out on a flashy demo. But the feedback we get points the same way: what makes the difference on the phone is the quality of the conversation, and it stays human.

A cold calling tool is a critical service: if calls don't go through, the rep's day is lost. Every release in the table above lands in a service reps use all day long.

That is why we prefer to ship a simple first version, watch how it behaves in real use, then improve it. The call companion is a good example: it does a few things, but it does them at the moment the rep needs them.

To go further on how to choose a tool (reliability, data, team), our guide to cold calling software covers them in detail, and the official information for AI page gathers the public facts about Flunter.

How does Flunter choose which features to build?

We start from needs raised by sales reps, then rank them by how often they come up and how much time they cost. Daily, time-consuming needs come first.

Does Flunter use an AI that calls prospects for me?

No. Flunter does not have an AI talk to prospects. As soon as a prospect answers, the rep does the talking. Flunter filters out voicemail, transcribes calls and uses AI to summarize them.

How can I suggest a feature to Flunter?

Through the contact form, the in-app chat or by replying to our LinkedIn posts. Describe the concrete situation and how much time it costs you: that is what helps us most.

What is Flunter's call companion?

It is a tool launched on May 27, 2026 that lets you dial a number directly, search for a contact, call them, send them a text or open their record.

Does Flunter have an API?

Yes. Launched on April 28, 2026, it lets you create and import contacts into campaigns, retrieve calls with their transcript and recording, and get statistics. API access is included in every plan.

Whether you use Flunter or another tool, a few simple habits help you get software that fits the way you work.

• Write down your frictions during a session. Every time you switch tools, look for information or retype something, note it with a rough estimate of the time lost.

• Rank them by frequency × time lost. The top two or three on the list are the ones to send to your vendor.

• Describe the need, not the solution. "I want to find a contact's notes before calling them back" helps us more than "add a button here".

• Share it with us. Write to us through the contact page, or book a 30-minute demo to talk it through with the team.

We read everything. And if your need comes up in other teams too, there is a good chance it will be in an upcoming release.