Cold outreach to AI & machine learning companies
The target buyer is a Product Manager at an AI & machine learning company, primarily measured on the successful deployment of new features in their LLM products. They respond to outreach that showcases a deep understanding of AI model performance metrics and how to optimize them, especially when faced with challenges in deployment timelines and resource allocation.
Written and maintained by the SoloRiff team, who build outbound software and run outbound with it. The figures below are our own: we measured 561 ai & machine learning company domains over public DNS on 2026-09-03. Last updated 2026-09-03.
What actually hurts
- deployment delays
- resource allocation issues
- model performance optimization
- feature integration challenges
The playbook
- 01
Identify recent LLM launches
Use platforms like Product Hunt or TechCrunch to find companies that have recently announced LLM product launches. Filter for announcements made in the last 30 days to ensure relevance. This helps you connect with teams excited about enhancing their offerings.
- 02
Analyze performance metrics
Research the performance metrics commonly discussed in AI forums, such as accuracy and latency of LLMs. Identify companies struggling with these metrics in their recent launches. This allows you to tailor your outreach to address specific pain points they may be facing.
- 03
Craft a message around optimization
Compose your email focusing on how your insights can help optimize their LLM's performance. Mention specific challenges like latency or accuracy that they might be experiencing, and propose a brief discussion to explore solutions. This makes your outreach relevant and timely.
- 04
Follow up after industry events
Monitor upcoming AI conferences or webinars and time your follow-ups for a week after these events. Reference discussions or insights shared during these events to demonstrate relevance. This increases the chances of a response as they may be reflecting on the topics covered.
- 05
Leverage case studies in outreach
Identify case studies or success stories that relate to the specific challenges faced by LLM product managers. Use these in your outreach to build credibility and illustrate potential outcomes. This gives them a clear vision of how your insights could benefit their projects.
Can ai & machine learning companies even receive your mail?
We measured the DNS of 561 ai & machine learning companies, sampled evenly across company sizes, to see how many authenticate their own email. It matters in both directions: a domain that does not authenticate is a domain whose team is used to mail going missing — and if your own domain is in the gap, yours is the mail going missing.
Publishes a DMARC policy of quarantine or reject — the domain actually asks receivers to act on failures.
The gap is the finding: 26.3% of 1–10-person companies against 86.2% of the largest — 60 points apart. Small companies are where the unauthenticated domains are, and small companies are who most cold outreach is aimed at.
Where does your domain sit?
Same three lookups, run live against your domain. Nothing is stored and nothing is emailed to you.
How this was measured
Company domains were sampled within employee-count bands, then queried over public DNS for SPF (TXT at the apex), DMARC (TXT at _dmarc.) and MX. One domain per company, deduplicated. Anyone can reproduce any single row of this with dig.
DKIM is deliberately absent. A DKIM key lives under a selector that cannot be enumerated from outside the domain, so a company with flawless DKIM under a name we did not guess would be counted as having none. That would measure our guess list, not the industry.
Domains with no MX record are excluded rather than counted as failures — parked and redirect-only domains would otherwise drag every figure down and make the result an artefact of the sample.
Sampled within bands, so this describes the population of companies, not of employees: a 10-person company counts once, as does a 10,000-person one. Each band is at least 60 domains; bands below that are not published. Figures are percentages of the sample, not of the whole industry, and we publish no company names — only counts.
Measured 2026-09-03. Re-measured monthly.
The full census across every industry — sortable, with sample sizes, free to reuse under CC BY 4.0.
Subject lines that fit
- Optimizing your LLM performance
- Insights on recent LLM launches
- Enhancing AI model efficiency
Questions
- Why do product managers in AI ignore cold outreach?
- Product managers in AI often ignore cold outreach due to the overwhelming volume of irrelevant messages. Many receive generic pitches that don't address their specific challenges, leading to a lack of engagement. To capture their attention, outreach needs to demonstrate a clear understanding of their unique pain points and industry trends.
- What challenges do AI product managers face with LLMs?
- AI product managers frequently deal with challenges such as integrating new features into existing systems, ensuring model performance meets user expectations, and managing deployment timelines. These issues can lead to frustration and delays, making targeted outreach that addresses these concerns particularly valuable.
- How can I make my outreach stand out to AI companies?
- To stand out, personalize your outreach by referencing specific LLM projects or recent launches. Highlight your understanding of the AI landscape and the unique challenges they face, such as model optimization or deployment hurdles. This targeted approach demonstrates that you’ve done your homework and respect their time.
- What timing is best for reaching out to AI product managers?
- The best timing for outreach is typically after they have launched a new product or feature. This is when they are most focused on performance metrics and may be open to discussions about optimization. Additionally, reaching out shortly after industry events can also yield positive responses as they reflect on new insights.
Or have it run itself
SoloRiff does every step above on its own — finds the companies, finds the people, writes each of them individually, and handles the replies. Drop your URL and watch it work before you sign up for anything.
Try it on your site