Most outbound dies on one decision: who to write to this week. Ami has already made it 17,707 times and kept score. Built on GPT-6 Astra and every one of those campaigns, it reads your website, picks the buyers who actually convert, and plans backwards from the number you need: leads, senders, weeks. If the list you asked for won't work, it says so before it builds anything. Approve the plan and it runs the outreach, watches every campaign, and fixes what slips. 19,854 booked meetings behind it.
Building AiSDR has made me rethink how much of sales can be solved by automating the tasks.
We started with prospect research, writing outreach, and handling replies. But I kept joining customer calls to help answer more basic questions: Who should we target? What should we offer? Why would someone care right now? What should we change when nobody responds?
I once paid an outbound agency $30,000 and got zero meetings in six months.
When we built AiSDR, I learned how easy it is to automate the emails while leaving the hardest decisions to the customer: who to contact, why now, and what to offer.
Ami uses GPT-6 Astra to plan, launch, and improve outbound sales campaigns.
For example, it turns LinkedIn engagement into outreach: identifies relevant buyers, researches their businesses, recommends an offer, and prepares personalized messages. After customer approval, Ami executes via email and LinkedIn, handles replies, and books meetings. A user request typically involves 4–5 tool calls that are all routed and executed by Astra.
We supply 27 industry playbooks built from 2.5 million emails, backed by experience from 17,707 historical campaigns. On every turn, we add relevant customer context, past decisions, and live campaign results.
As replies arrive, Astra helps decide what to change: audience, offer, messaging, or whether to stop. Ami executes within the customer’s approved scope.
Hi Product Hunt. Yuriy here, co-founder and CEO of AiSDR, 2x YC alumni.
I once paid an outbound agency $30,000 and got zero meetings in 6 months.
Then I built an AI SDR, sold it to 1,153 companies, and watched 80% of them get the same result, but faster and cheaper. The emails were fine. Nobody, including our software, solved the core problems - who should get them and why should they care to respond.
AI SDR industry overpromised and underdelivered and people hated it.
Ami is the thing I wish I could hand to every customer who trusted us in 2023 and got a fast way to fail.
Here's what it looks like.
You give it your website. It comes back with three campaign ideas, each with a signal behind it and a reason it thinks this one is likely to deliver results.
That answer is not yet another AI slop. It's rooted in 17,707 campaigns we've run across 27 industries that booked 19,854 qualified meetings.
You say "actually, I want every salesperson in Europe." It tells you that the audience has booked zero meetings across all campaigns we've run, and hands you the version that will.
You say "I need 100 demos by October." It gives you the leads, the infrastructure, and the week the plan breaks if one of them slips.
You approve. It builds campaigns and runs them across email and LinkedIn. When a campaign starts dying it tells you which of three things broke (audience, message, or timing), proposes a fix, fixes it for you and remembers what you chose. It also suggests new ideas as it learns more about you and your business.
It takes care of all the boring stuff - email infrastructure, deliverability, protecting your main domain and LinkedIn account from being blocked, reply classification and handling and dutifully reports all of it's activity into your CRM.
Why this exists now and not two years ago: GPT-6 Astra.
Models could write outreach for a while. What they couldn't do was hold a multi-step plan, check it against tens of thousands of real outcomes, and change course mid-flight without losing the thread. Astra, with it's 96.3% on million-token recall benchmark can. That's the reasoning layer; our three years of campaign data is what it reasons against.
Numbers from three months of beta: $697M in pipeline influenced. And this one, a Solo user's first campaign: 18% of contacts replied, 47% of those replies were positive. I asked to see the dashboard before I'd say it out loud.
Who it's for: first-time founders, first sales hires, teams that need to get started with outbound to grow pipeline, but where nobody has done outbound before. If you already have a great SDR team, you probably don't need us.
Try it now. Point Ami at your website and you'll have your first campaign plan in about 15 minutes. No credit card, no demo.
Product Hunt code AMIHUNT50 takes 50% off Solo for 3 months.
app.aisdr.com/signup
Then come back and tell me what it got wrong.
I want this to be the AI sales agent you've been waiting for, so if the targeting is off or messaging reads like AI slop or UX is bad - I want to hear from you and see screenshots and I promise - we will improve.
I'll be in a thread all day and if you want to connect 1-1 to share feedback - you can grab a slot with me here: https://calendly.com/aisdr-intro...
Report
💎 Pixel perfection
@yuriy_zaremba I love that it suggests fixes while the campaign is running instead of waiting for a depressing post-mortem, such a nice way to so solving congrats for launch team🙌
@dimneo, thank you! Astra made all the difference in terms of performance
Report
@yuriy_zaremba nice launch... the idea of the testing campaign ideas against the real promotion data is the interesting, and also the audience targeting part caught my attentions.
how quickly does it the usually detect that somethings is wrong???
@vipul_kumar1280, 2 weeks is the first check and 3 weeks is the final decision. Needs some volume to make a good decision
Report
@yuriy_zaremba Congrats on the launch! Opening with "$30k for zero meetings" and then admitting 80% of your own customers failed is a rare way to start a launch - made me trust the rest of the post more!
You've named first-time Founders as your buyer, and they're also the people least equipped to judge the two things that carry the most risk here. So from that seat, how does Ami connect to LinkedIn, and if an account gets restricted, what's the path back? And when I approve the plan, am I approving the targeting, or the actual messages before they go out?
Report
Looks promising! Congrats on launch! What tools do you offer to control the quality of leads brought with your outreach? I see stats as Total response rate and Positive response rate, can you elaborate?
Report
Maker
@liya_bondarenko thanks! We usually measure success against Positive reply rate and Meetings booked count, as well as Influenced pipeline. With outreach it can easily turn into an associated conversion story - someone clicks the outreach and then comes back in a month directly, so influenced pipeline is a good way to track the impact, especially with products that have long decision-making cycles
@liya_bondarenko, there are 3 filters: (1) lead list creation - it's agentic, not based on filters, so the goal is to discover people, who have a pain you can solve, not just a random list; (2) the messaging profile - you configure (with Ami) what is the offer and whether AI should qualify incoming leads, (3) iteration - as positive replies start coming in - you can fine tune both messaging and lead list to make sure you're getting qualified leads.
On the metrics side - response rate is everyone who responded, including saying "no/unsubscribe, etc.", while positive response is "yes, let's talk".
We also track pipeline influenced and $$$ closed won, because we are so focused on results
Report
Hi guys!
the product looks very interesting (and to be covering the biggest gap almost every "I am building Xxx" product has)!
I am not sure I understand where the contact data are coming from and if from my side, how do you handle any data from my side? If it is from the model seed, how do you handle quality of the contacts?
Looks awesome!
Report
Maker
@foobar_beer thanks Jakub! The data is coming from a waterfall of over a dozen data vendors we work with including public web data scraping for live signals. We also triple-verify every contact for potential bounces, which is how we achieve a system-wide bounce rate <1%
Report
@alina_karnaukh Hi Alina, thanks for the answer - so can those contacts be added to my list as (ideally double) opt-in contacts?
@foobar_beer, we have multiple vendors under the hood, including live scraping and 2 bounce providers. Data is the best you can get in the market. We've iterated on the setup for 3 years
Report
How does Ami know that the list I asked for won't work. Where is her judgment coming from?
@maryna_synychenko, great question. Ami is trained on almost 18k campaigns we've run over 3 years AND on top of it scans what's latest and greatest in GTM every week, so it has experience and acts as an expert which is why it can push back on bad ideas
Report
Hey team! One thing that drives me crazy with AI is how agreeable it is. I can give it a pretty questionable idea and it’ll still find a way to tell me it’s great 😂
Can Ami actually tell me when I’m onto something bad?
@new_user_112af36f66, yeah. It's very non-agreeable. It focuses only on things that it knows will work. We specifically trained it to be like that, because many customers that we've worked with - were doing what they felt was right and failing
Report
Congrats on the launch. I’m very interested to try it out because I’ve trialed a few outreach platforms lately and haven’t had much success. I put that down to the fact that I know very little about it, so I really like the built-in fix/recommendation feature. If it sends emails, how does it warm up the list? One thing I’ve learnt is about low réponses rates from cold outreach.
@you_x_you_i, list is warmer, because Ami always suggest a lead list that starts with a pain signal (new pricing page/new hire/M&A) or a social signal (post likes, profile views) and so people are already "aware of the problem" and are more likely to respond. On top of that we take care of all the infrastructure issues on email deliverability and LinkedIn side
Report
@yuriy_zaremba interesting about the LinkedIn point. Does it automate the DM?
Ami AI
Ami AI
@yuriy_zaremba I love that it suggests fixes while the campaign is running instead of waiting for a depressing post-mortem, such a nice way to so solving congrats for launch team🙌
Ami AI
@vikramp7470, thank you! Yes. This is key. It checks performance weekly and suggests to make adjustments to make sure campaign works
Ami AI
@tehreem_fatima5, all the greatest changes are ahead
@yuriy_zaremba Congrats on the launch! The positioning is sweet! Also, I like the adaptive behaviour of Ami, using Astra makes sense!
Ami AI
@dimneo, thank you! Astra made all the difference in terms of performance
@yuriy_zaremba nice launch... the idea of the testing campaign ideas against the real promotion data is the interesting, and also the audience targeting part caught my attentions.
how quickly does it the usually detect that somethings is wrong???
Ami AI
@vipul_kumar1280, 2 weeks is the first check and 3 weeks is the final decision. Needs some volume to make a good decision
@yuriy_zaremba Congrats on the launch! Opening with "$30k for zero meetings" and then admitting 80% of your own customers failed is a rare way to start a launch - made me trust the rest of the post more!
You've named first-time Founders as your buyer, and they're also the people least equipped to judge the two things that carry the most risk here. So from that seat, how does Ami connect to LinkedIn, and if an account gets restricted, what's the path back? And when I approve the plan, am I approving the targeting, or the actual messages before they go out?
Looks promising! Congrats on launch! What tools do you offer to control the quality of leads brought with your outreach? I see stats as Total response rate and Positive response rate, can you elaborate?
@liya_bondarenko thanks! We usually measure success against Positive reply rate and Meetings booked count, as well as Influenced pipeline. With outreach it can easily turn into an associated conversion story - someone clicks the outreach and then comes back in a month directly, so influenced pipeline is a good way to track the impact, especially with products that have long decision-making cycles
Ami AI
@liya_bondarenko, there are 3 filters: (1) lead list creation - it's agentic, not based on filters, so the goal is to discover people, who have a pain you can solve, not just a random list; (2) the messaging profile - you configure (with Ami) what is the offer and whether AI should qualify incoming leads, (3) iteration - as positive replies start coming in - you can fine tune both messaging and lead list to make sure you're getting qualified leads.
On the metrics side - response rate is everyone who responded, including saying "no/unsubscribe, etc.", while positive response is "yes, let's talk".
We also track pipeline influenced and $$$ closed won, because we are so focused on results
Hi guys!
the product looks very interesting (and to be covering the biggest gap almost every "I am building Xxx" product has)!
I am not sure I understand where the contact data are coming from and if from my side, how do you handle any data from my side? If it is from the model seed, how do you handle quality of the contacts?
Looks awesome!
@foobar_beer thanks Jakub! The data is coming from a waterfall of over a dozen data vendors we work with including public web data scraping for live signals. We also triple-verify every contact for potential bounces, which is how we achieve a system-wide bounce rate <1%
@alina_karnaukh Hi Alina, thanks for the answer - so can those contacts be added to my list as (ideally double) opt-in contacts?
Ami AI
@foobar_beer, yes, they can be added to your list. We use only GDPR-compliant vendors
Ami AI
@foobar_beer, we have multiple vendors under the hood, including live scraping and 2 bounce providers. Data is the best you can get in the market. We've iterated on the setup for 3 years
How does Ami know that the list I asked for won't work. Where is her judgment coming from?
Ami AI
@maryna_synychenko, great question. Ami is trained on almost 18k campaigns we've run over 3 years AND on top of it scans what's latest and greatest in GTM every week, so it has experience and acts as an expert which is why it can push back on bad ideas
Hey team! One thing that drives me crazy with AI is how agreeable it is. I can give it a pretty questionable idea and it’ll still find a way to tell me it’s great 😂
Can Ami actually tell me when I’m onto something bad?
Ami AI
@new_user_112af36f66, yeah. It's very non-agreeable. It focuses only on things that it knows will work. We specifically trained it to be like that, because many customers that we've worked with - were doing what they felt was right and failing
Congrats on the launch. I’m very interested to try it out because I’ve trialed a few outreach platforms lately and haven’t had much success. I put that down to the fact that I know very little about it, so I really like the built-in fix/recommendation feature. If it sends emails, how does it warm up the list? One thing I’ve learnt is about low réponses rates from cold outreach.
Ami AI
@you_x_you_i, list is warmer, because Ami always suggest a lead list that starts with a pain signal (new pricing page/new hire/M&A) or a social signal (post likes, profile views) and so people are already "aware of the problem" and are more likely to respond. On top of that we take care of all the infrastructure issues on email deliverability and LinkedIn side
@yuriy_zaremba interesting about the LinkedIn point. Does it automate the DM?
Ami AI
@you_x_you_i, yes it does, as well as voice notes (AI voice), likes, InMails