14.ai Dispatch #23
Sushi and customer support walk into a bar.
AI agents handling TikTok Shop refunds.
Protecting your brand on Reddit.
Basically: Social commerce and reputation. Because that’s all that seems to matter lately.
Event recap: AI agents for e-commerce & DTC



Between sushi platters and games of telestrations (essentially telephone pictures), we built the next great e-commerce agent.
We gathered founders, operators, and support leads who are sick of answering “wheres my order” and legacy ticketing systems.
Overall, major success. We couldn’t fit everyone this time around, but will be doing more soon (if you follow this newsletter, you know we do these like every day).
Podcast recap: Customer feedback (and why you should fear Reddit)
There’s probably nothing worse than a viral TikTok dragging your product.
You’ll either:
1️⃣ Catch a class-action lawsuit, or
2️⃣ Watch your customers vanish.
Neither option’s great.
We talked with Alex J. Hong from Syncly (YC W23), who’s trying to make sure that doesn’t happen.
Syncly tracks and centralizes customer feedback across every platform; because seriously, who submits support tickets anymore? I’d rather complain in a tweet.
Here’s what stuck with us (and yes, Michael posted this on LinkedIn, but we’ll save you the scroll):
1. CSAT and NPS are lagging indicators.
By the time someone leaves a 3/10, it’s already too late. The focus is sentiment tracking: catching frustration (anywhere) early and resolving it before it snowballs.
2. The next wave of CX feels more like cybersecurity.
Detect spikes fast, identify if it’s misinformation, product confusion, or genuine issue. Contain and respond before it trends. The faster you close the loop, the fewer TikTok “horror stories” you’ll have to clean up.
3. AI
Listens, categorizes, and clusters issues automatically (pricing, delivery, UX, etc), and routes them to the right people, so teams can focus solely on resolutions.
Reputation management = latency management.
So yeah. Watch your back. And your reddit.
Thanks, Alex 🙏
Next week: AI for short-form content with Bluma
Short-form video is the new storefront.
The question is: who’s running it... you, or an AI?
Join Alisa Wu, CEO and co-founder at Bluma (YC F25), for a live session on how AI is transforming how brands create, scale, and personalize short-form content.
Basically: low effort, high output, and hundreds of on-trend videos.
Please shoot us any questions you’d like us to ask (on the AI short form video space in general)
We’ll cover:
Capitalizing on trends at scale
Balancing automation, authenticity, and creative control
The future of AI-generated short form (and yes, hopefully more of those “Sora” videos everyone’s reposting on Facebook)
Perfect for: Founders, marketers, and creative teams at e-commerce and consumer brands.
Come through!
📆 Wednesday, November 12 @ 9 AM PT
Insane cold emails
Maybe the best one yet. And yes, great name.
In case you missed it
Cursor for stepping: YC F25’s take on “going for a walk,” powered by steps per token.
OneText origin story: When your phone number becomes your wallet.
Agentic lessons: From docs drift to hybrid retrieval, the new rules for AI support ops.
Taming complexity: Matthew explains why context > prompts.
YC immigration: F25 batch made it through customs. Group partner: Jerome Powell.
Red Room > Zoom Room: Barry’s streak still alive.
Whiteboard goals: Because “Exercise 3x” belongs right next to “Ship features” and “Close loops.”
Cool jobs at 14.ai (and friends)
🛠️ Founding Engineer @ 14.ai (San Francisco)
⌨️ Support Executive @ Landeed (Remote)
💻Customer Support Specialist @ VetRec (Seattle, WA)
What we’re reading (so you don’t have to)
Marie: The 7 Habits of Highly Effective People, Stephen R. Covey
Matthew: Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models, Zhang et al.
Maya: In a First, AI Models Analyze Language As Well As a Human Expert, Steve Nadis
Michael: Optimism is associated with exceptional longevity in 2 epidemiologic cohorts of men and women, Lee et al., 2019
Ugo: Patterns for building AI agents, Sam Bhagwat
Until next time: may your refunds resolve fast, your TikTok comments stay kind, and your agents never hallucinate.









