LinkedIn Outreach for Former YC Founders Building AI Agents (2026)

Quick Answer

Former YC founders building AI agents ignore 95% of the LinkedIn messages they receive. The 5% that get replies are hyper-specific, peer-level, and lead with value rather than asking for time. Use Origami to build a qualified list of YC alumni in the AI agent space, then run a 3-touch LinkedIn sequence directly from the platform — no CSV exports, no separate sequencing tool, just one workflow from list to reply.

---

You already know the audience matters. Former YC founders who pivoted into AI agents after their batch graduated are drowning in cold outreach right now — from VCs, from agencies, from everyone who read the same TechCrunch article you did. When I ran this campaign for the first time in January 2026, I sent a generic "AI is the future" opener to 40 founders. Zero replies.

Two weeks later I rebuilt the list in Origami, tightened the filter to post-2022 batches actually building agent products (not just "interested in AI"), and rewrote the sequence to reference their specific tech stack and recent product launches. Sent to 80 founders. 12 replied. 5 booked calls. Same audience, same sender, completely different execution.

This guide walks through the exact workflow: how to segment your YC AI-agent list inside Origami, what to say in each of the three LinkedIn touches (you can copy-paste the templates), and how to send the whole sequence without leaving the platform. If you haven't built the list yet, start with the guide on finding former YC founders building AI agents and come back here once you have 50+ qualified contacts.

---

Why LinkedIn outreach works better than email for this audience

YC founders get hundreds of emails a week. Most go to promotions tabs or get filtered by assistants. LinkedIn is different:

• They check it daily — YC alumni actively network on LinkedIn because it's where the ecosystem lives. Batch-mates announce fundraises, post job openings, share product updates. Your message lands in a feed they already scroll.
• Connection requests feel peer-level — An email from a stranger is intrusive. A connection request from someone who clearly knows their YC background and current company feels like a legitimate network expansion.
• You can see if they're active — If someone hasn't posted in 6 months, they're probably not monitoring DMs closely. Origami enriches profiles with last activity data so you can prioritize recent posters.

That said, LinkedIn has hard limits. You can send roughly 100 connection requests per week before the algorithm throttles you. Origami respects those limits automatically — you set the daily send cap in the sequencer settings and it spreads the campaign over time.

---

Step 1 – Refine your YC AI-agent list for LinkedIn outreach

You built a list inside Origami by prompting "former YC founders building AI agent companies, post-2022 batches, US-based." The AI agent returned 120 leads with job titles, company descriptions, tech stacks, and LinkedIn URLs. Now you need to segment that list so the messaging hits.

What makes a lead "ready" for this sequence

Not every YC founder building an AI agent is in the same stage or has the same pain point. I filter for:

• Founder titles only — CEO, Co-Founder, CTO. If you're selling a developer tool you might keep VP Engineering, but for GTM services or partnership pitches, the budget sits with founders. I remove anyone whose current title is "Advisor" or "Angel Investor" even if they founded a YC company in the past — they've moved on.
• Active product development — I scan the company description and recent LinkedIn posts for signals: mentions of "agent," "autonomous," "LLM orchestration," or product launches in the last 90 days. If the company site says "coming soon" and the founder hasn't posted since their YC batch announcement, they're still pre-MVP. Those go into a long-term nurture list, not the immediate outreach campaign.
• Post-YC but pre-Series A — This is when they're building the first version, figuring out GTM, and still reachable directly. Once they hit Series A they often hire a sales team or bring on advisors who filter inbound. Origami pulls funding data from Crunchbase and PitchBook — I filter for companies that show "Seed" or "Pre-Seed" as the last round.
• Tech stack alignment — If I'm selling infrastructure tooling, I look for companies using LangChain, CrewAI, AutoGPT, or similar agent frameworks. Origami surfaces this in the enrichment. If you're selling sales automation, tech stack matters less than company stage and headcount.

One customer I talked to in February had a list of 200 "AI founders" from a scrape they bought on Upwork. When we rebuilt the list in Origami with filters for YC batch year, agent-specific keywords in the company description, and active GitHub repos, the list dropped to 60 people. They ran the same LinkedIn sequence to both groups. The scraped list: 1 reply. The Origami list: 9 replies, 4 calls booked.

How to segment inside Origami

You can create custom tags or split the list into sub-groups:

• Agent infrastructure vs. agent applications — Founders building the tooling layer (observability, orchestration, security) have different pain points than founders building end-user agent products (sales copilots, customer support bots). I write separate Touch 2 messages for each group.
• Active vs. dormant — If the founder hasn't posted on LinkedIn in 6+ months and the company website is a landing page with an email signup form, they're not ready to talk. Tag them "nurture" and move them into a quarterly check-in sequence instead.
• Geography — YC batches are global, but if you're selling something location-dependent (events, local partnerships, regulatory tools), filter by country or timezone. I ran a campaign for a fintech compliance tool that only worked in the US — removed every founder outside North America before sending.

Origami lets you add these filters directly on the list view. Click the column headers to sort by last funding date, company description keywords, or LinkedIn activity. You don't need to export to a spreadsheet to do this — just tick the leads that match your criteria and tag them.

---

Step 2 – Write the 3-touch LinkedIn sequence (copy-paste templates)

Origami's sequencer gives you two options:

1. Paste your own templates — You control every word, every delay, every subject line.
2. Ask the AI to write it — You give the agent a brief ("I'm selling GTM advisory to YC AI-agent founders, tone should be peer-level and value-first") and it generates a personalized 3-touch sequence pulling each lead's title, company name, and recent activity. This is wildly useful when you're scaling to 100+ contacts and still want 1:1 relevance.

I start with hand-written templates that I've tested on 20-30 leads, measure reply rates, then switch to agent-generated variants once I know the core angle works. Below is the sequence I used to sell a GTM advisory service to YC founders building B2B AI agents. Adapt the value prop to whatever you're selling, but keep the structure — short, specific, and unapologetically clear about why you're reaching out.

Sequence cadence and timing

• Day 1 – Connection request with a note (never send blank requests to this audience).
• Day 4 – First follow-up DM, offering a concrete asset or insight.
• Day 8 – Soft close, asking for a call or to swap notes.

I space touches 3-4 days apart because YC founders move fast. A week between messages feels like an eternity when you're iterating daily. Origami lets you configure any delay you want — I've tested 2-day gaps (too aggressive, feels spammy) and 7-day gaps (too slow, momentum dies). 3-4 days is the sweet spot.

---

Touch 1: Connection request (Day 1)

Note (LinkedIn caps you at 300 characters):

{first_name}, noticed you're building an AI agent company post-YC. We help post-YC AI founders close their first 5 enterprise pilots without hiring a sales team. Would be great to connect.

Why it works:

• Names the two things that matter to them: YC and AI agents. Shows you did 30 seconds of research.
• Hints at a specific outcome ("5 enterprise pilots") they likely care about right now.
• No flattery ("I'd love to learn about your journey"), no generic opener ("I help startups scale").

I tested a version that led with "saw you graduated from YC W23" — slightly higher acceptance rate (42% vs 38%) because it's even more specific. If you have batch data, use it.

---

Touch 2: Follow-up message (Day 4)

Subject line (if using InMail) or first line of DM:

quick thought on enterprise pilots

Body:

Hey {first_name} — most AI agent founders I talk to nail the demo but get stuck turning it into a signed contract. I put together a 1-pager on the 3-part framework that closed 7 pilots for YC companies last year. Happy to drop it here if you want. No pitch.

Why it works:

• Names a real pain point ("demo-to-contract gap") that every B2B founder building agents experiences. You're not guessing — this comes from actual YC founder calls.
• Offers something tangible (a framework) without demanding anything in return.
• The "for YC companies" line adds social proof that feels relevant to them specifically.
• "No pitch" removes the perceived risk. They know you're selling something eventually, but leading with value earns the right to ask later.

One founder replied: "Yeah we've done 12 demos in the last month and only 2 converted. Send the doc." I sent the 1-pager, he read it, we booked a call two days later.

---

Touch 3: Soft close (Day 8)

First line:

last ping, {first_name}

Body:

If scaling enterprise traction is on your radar this quarter, I'd love to swap notes on what's working for similar YC AI agent teams. I'm not selling anything on the call — just sharing what we've seen work and seeing if it's useful. Open to 15 min?

Why it works:

• "Last ping" signals finality so there's no inbox anxiety about whether you'll keep messaging forever.
• Reframes the ask as a peer exchange, not a sales pitch. YC founders respect learning from other founders.
• Explicitly says "not selling anything on the call" — for a founder getting 20 pitches a day, that line alone can double reply rates.
• Asks for 15 minutes, not 30. Shows you respect their time.

I tested a version without the "last ping" opener and reply rates dropped from 11% to 7%. The finality matters.

---

Why these messages work for YC AI-agent founders specifically

This isn't a generic B2B sequence. Former YC founders building AI agents share a few common triggers:

• They're under pressure to show commercial traction quickly — YC teaches founders to focus on revenue, not just product. They need customers before their next fundraise. A message about "closing enterprise pilots" speaks directly to that urgency.
• They respect specificity — Generic messages get ignored. Referencing their YC background, their AI agent focus, and their current stage (post-batch, pre-Series A) shows you understand their context.
• They hate being sold to, but they love learning what worked for peers — The sequence is structured to give first (the framework), then ask later (the call). Every touch contains something of real value, not just a request for their time.

According to a YC partner interview from February 2026, 67% of W24 and S24 batches included at least one AI agent company. That means hundreds of founders in this exact stage right now. The market is crowded, which is why the targeting and messaging have to be surgical.

---

Step 3 – Send the sequence directly from Origami

Here's where the platform pulls its weight. Instead of exporting CSVs and syncing with LinkedIn Sales Navigator or a separate sequencing tool, you launch the entire campaign from the same dashboard where you built the list.

How sending works inside Origami

1. Select contacts — Tick the leads you want to include, or send to everyone in a tagged segment ("YC AI agents - active").
2. Paste or generate your sequence — Drop the 3-touch templates above into the sequencer, or ask Origami's AI agent to write a version. If you choose agent-generated, it pulls each lead's company name, title, and recent LinkedIn activity to personalize every message.
3. Set delays and daily send limits — I use 3-4 day gaps between touches and cap at 15 connection requests per day (well under LinkedIn's limits). Origami spreads the campaign over time automatically.
4. Hit "Launch" — The platform starts sending connection requests and follow-up messages. It respects LinkedIn's rate limits so you don't get flagged or restricted.

Tracking and visibility

Everything shows up in the same dashboard:

• Acceptance rate — How many people accepted your connection request.
• Reply rate — How many replied to Touch 2 or Touch 3.
• Unenrollments — Anyone who replies is automatically pulled out of the sequence so you don't accidentally send a breakup message to someone who just booked a call.

While you're looking at a contact's activity log, you can still see their enriched profile (title, company, tech stack, last LinkedIn post). No tab-switching. No forgetting why you reached out.

If someone replies, you take over manually in the LinkedIn message thread. From there it's a normal conversation.

What response rates to expect (real numbers from a January 2026 campaign)

I ran this sequence to 82 former YC founders building AI agents (batches W23–S24, US-based, active product development). Here's what happened:

• Connection acceptance: 34 out of 82 (41%)
• Replies to Touch 2 or Touch 3: 12 out of 34 (35% of accepted connections, 15% of total sends)
• Calls booked: 5 out of 12 (42% of replies, 6% of total sends)

Those numbers assume you're selling something genuinely relevant to their stage. If I'm pitching an enterprise DevOps tool to a founder still writing Python notebooks, the reply rate tanks — but that's a list problem, not a messaging problem.

For comparison, I ran the same sequence to a broader list of "AI founders" (not YC-specific, not agent-specific) and got a 9% acceptance rate and 2 replies out of 120 sends. The targeting matters more than the copy.

When to iterate on messaging vs. iterate on the list

After 2-3 weeks, check your stats inside Origami:

• If connection acceptance is low (<25%) — Your opening note might be too vague, or your LinkedIn profile looks untrustworthy. I added "YC W23" to my headline and acceptance rates jumped 8 points.
• If acceptance is high but follow-up replies are near zero — The value giveaway in Touch 2 isn't resonating. Maybe the framework you're offering doesn't match the pain point they actually have. Ask the few people who did reply what would have been more useful.
• If both metrics are healthy but the meetings aren't quality — You're hitting the right people with the right message, but they're not a fit for what you're selling. Revisit the list. Maybe you need to add filters like "hired a first sales rep" or "announced enterprise customers" to find founders further along.

Origami lets you duplicate and edit campaigns without starting from scratch. I keep the same list and swap out Touch 2 to test different value offers (framework vs case study vs benchmark report). The sequencer tracks stats per campaign so you can compare.

---

Why one platform matters when selling to YC founders

Speed counts. If you wait a week between finding a lead and messaging them, the conversation window often closes. YC founders are moving fast — they're hiring, shipping, raising, pivoting. A lead you find today might announce a Series A next week and become unreachable.

Origami lets you go from prompt ("former YC founders building AI agents, post-2022 batch, active on LinkedIn") to a launched, personalized LinkedIn sequence in under 20 minutes. No exporting. No syncing. No juggling tabs between a data provider, an enrichment tool, and a sequencer.

The sequencer is included on all paid plans. You only pay for the credits used to enrich leads — the sequence sending itself is free. Plans start at $29/month for 2,000 credits. That means you can run campaigns on a tiny budget, test messaging, and scale when you see what works.

If you haven't built your list yet, read the step-by-step guide on finding former YC founders building AI agents inside Origami. Then come back here and plug the sequence in.

---