Decipher AI (YC W24)

I was speaking with a customer who was STRUGGLING to extract value from Fullstory's session replays. I get it. While Fullstory attempts to highlight interesting moments such as dead clicks, its pre-LLM design means the detection is quite naive. As a result, you still need to watch each replay manually to identify actual issues. Just look at this example from Fullstory (on our user sessions): 95% of the dead clicks it highlights are false positives. Same with rage clicks and other signals. But this doesn't have to be the case. Next-generation session replay platforms like Decipher AI (YC W24) use vision language model agents to truly understand what's happening in your sessions and automate key workflows, such as finding issues and analyzing customer behavior. There's simply no reason to pay a session replay platform six figures just to store replays you rarely watch. The next generation is here.

Customer churn can stem from many causes—but buggy software is a frequent offender. So, how do you catch issues and prioritize them before they escalate? Meet Decipher AI (YC W24)'s new "Had An Issue" replays filter. With one click, see every customer replay where our AI agent watched the replay and flagged a bug. Plus, you can even filter by customer to quickly determine if your top accounts are encountering problems.

🚀 Exciting News! 🚀 We’re thrilled to introduce our new Linear integration on Decipher AI (YC W24). Now, our AI agent watches your session replays and feeds discovered customer issues directly into Linear. Here’s how it works: 1️⃣ Detection: Our AI agent continuously monitors your session replays with visual LMs to spot any issues. 2️⃣ Alert: The moment a problem is detected, you’ll receive a real-time Slack alert. 3️⃣ Action: A detailed Linear ticket can be automatically created—complete with reproduction steps. Plus, some of our more AI-advanced customers are having great success by tagging in Devin, so the fix is already under investigation in the background. This streamlined process ensures that you discover and prioritize important issues before they cause an NPS hit or churn, allowing your team to focus on building value-add features.

Our AI agent analyzed thousands of session replays for a new customer and within hours it found a critical issue — many users weren’t able to log in via one of their OAuth providers 🚨 Here’s what happened: 🤖 Analysis: Once integrated, our AI agent started watching every session replay for every user on our customer's product. 🔍 Immediate Detection: Within hours, our agent flagged the login failures by analyzing hundreds of hours of session replays. 📊 Context-Rich Insights: The agent provided detailed session replay timelines and provided a list of which users were impacted. ⚡ Actionable Data: It identified relevant code-level information to assist engineers in debugging, including the response from a related network request and all associated logs. At Decipher AI (YC W24), our agent and advanced vision LMs do the heavy lifting, watching every session replay so you don’t have to. We love finding new issues for customers, so ping me and we'll get our AI agent analyzing your replays ASAP!

15 interesting AI agents to run marketing on autopilot ⚡ 1. PolyAI > Conversational assistants 2. ElevenLabs > Build AI Agents that speak 3. Coframe > Automate website growth 24/7 4. Jason AI > Outbound sales AI agent 5. Rep.ai > Live website conversations 6. quso.ai > All-in-one social media AI 7. Decipher AI (YC W24) > Agent for session replays 8. Keak > AI agent that improves website 9. Shown > Let AI run your ads 10. CrewAI > Multi-agent platform 11. SEO bot > Fully autonomous SEO Robot 12. Averi > AI marketing manager 13. MindPal > Build your AI workforce 14. beam > Hire self-learning AI agents 15. Argil (YC S24) > Generate engaging AI videos - - - - - - - - - - 🖖 Enjoy and follow me to stay updated on SaaS, AI, trends & new tools. 🔖 Bookmark to check these tools later.

I speak to tons of YC founders and I constantly hear the same story. The product dashboard shows 10K daily active users, and your team ships new features every sprint. But here's the uncomfortable truth: you probably have no idea which features are actually being used, and you're wasting precious engineering time. 😱 Consider this common scenario: A project management tool has grown to include: - Task management - Time tracking - Document collaboration - Sprint planning - and ten other features The product team celebrates DAU growth, customers request new features, and engineering keeps shipping. But dig deeper and you'll find: ❌ 70% of users only use basic task lists ⚙️ Most "power features" sit unused 🛠️ Teams create workarounds instead of using purpose-built features What you actually need to know for EVERY feature: 🔍 Basic Adoption - Which features are being used vs. completely untouched? - How many users return after their first try? - What's the adoption trend over time? 📈 Adoption Depth - Who exactly is using it and are they getting value? - Did they just click or complete the core workflow? - Are they using all capabilities or just the basics? 🗺️ Usage Patterns - What steps do users take before and after? - Where do they get stuck? - Which user segments get the most value? Why this matters: ✅ Product teams can finally focus resources on what works and fix what doesn't. 🧑🤝🧑 Customer Success can guide users to valuable features. 📊 Sales can show prospects exactly how similar companies succeed with your product. The results: 💔 Lower churn and higher NPS ⚡ Faster feature adoption rates 🎯 Confident product decisions backed by user behavior Here's the thing: Traditional analytics take forever to get even basic adoption insights. You don't want to waste engineering time instrumenting everything—you need to actually see how users interact with each feature. Luckily, AI’s got your back! At Decipher AI (YC W24), we empower teams to uncover how their features are truly being used, leveraging our vision LM agents to analyze thousands of hours of session replays—so you don’t have to.

Here is everything that happened in AI Agents this week 🧵 (save for later) 1/ Basis raised $34m to accelerate their goal of equipping every accountant with a team of AI agents https://lnkd.in/g7wDTUCW 2/ Fetch.ai wrote an article about multi-agent session tracking with uAgents and AgentOps from Agency. It goes over how to monitor, analyze, and optimize interactions between AI agents https://lnkd.in/gwUqmR9C 3/ Satya Nadella describes SaaS apps as nothing more than a crud database with some business logic, but once the business logic moves to AI agents, SaaS is over https://lnkd.in/gp5jEJTq 4/ Hamming AI raised $3.8M in seed funding https://lnkd.in/g-_TpYeX 5/ One of Agency's partners CrewAI released “Multi AI Agents in Production” https://lnkd.in/gaSvufR9 6/ Huge new “HTTP” column update to Cognosys! - This allows you to hit any webhooks with whatever data you want, essentially creating a way to sync your data with any CRM or any tool with a webhook https://lnkd.in/gndfUYfT 7/ Google Cloud gathered 321 examples of Generative AI having a real world impact across industries https://lnkd.in/g3ukfacZ 8/ “2025 will be the year of agentic systems.” - Alex Albert  from Anthropic https://lnkd.in/g4QekBSZ 9/ Chatbase announced AI Actions that turns your chatbot into an AI Agent that can do things like upgrade subscription, track shipment, book appointment, and more https://lnkd.in/gTsh2acs 10/ Stanford AI researcher Joon Sung Park + others created 1,000+ AI agents that perfectly mirror real humans https://lnkd.in/gS4vdGBw 11/ John Rush wants to replace 90% of support and operations in 2025. https://lnkd.in/giaAf_Kv 12/ Safwaan Khan made an AI Agent that can create Google Ads https://lnkd.in/gyBvdCRn 13/ Siva Surendira shared his experience getting to 100x productivity Wispr Flow for voice dictations, Granola for meeting transcripts, & Lyzr AI for copilot agents https://lnkd.in/gfmj9BHd 14/ Decipher AI (YC W24) helps you spot bugs, understand feature usage, and answer your questions in real-time. https://lnkd.in/gtkHfsVj 15/ PySpur: 100% open source drag, drop, & deploy to API in one click with NO CODE. https://lnkd.in/gVtq2erh 16/ Thomas Dohmke announced that “Copilot extensions are third-party agents designed for everyday tasks." https://lnkd.in/gX35tkyu 17/ LangChain shared their State of AI 2024 report https://lnkd.in/gNBCMDxG 18/ Akshay Pachaar created an agentic-workflow to automatically write and publish content https://lnkd.in/guKqCnUC 19/ SEO agent using AgentStack from Agency https://lnkd.in/gtJhXb5v 20/ Wayne Hamadi shared his article titled “You don't understand AI agents.” https://lnkd.in/g6aj64dW 21/ AG2 agents can talk to you in real time now! Built-in Twilio integration + others https://lnkd.in/gHtnYqW9 22/ Did I miss anything? Share below. Liked this thread? Share it with your friends/team. Follow Adam Silverman (Hiring) and Agency for everything AI agent related. Last weeks thread: https://lnkd.in/gXheQ2qQ

Decipher AI (YC W24) has become a part of my day! Best part is zero switching cost if you are already using Sentry! Then you will know exactly how to reproduce an error and possible way to fix it. Congrats Rohan D. and Michael Rosenfield 💪. You guys are extremely supportive ❤️❤️

Decipher AI (YC W24) offers an AI agent that watches thousands of session replays for you, enabling teams to uncover customer issues, track feature usage, and answer product questions—all in real-time. They’ve already analyzed millions of session replays for companies of all sizes. Traditionally, teams spend countless hours scrubbing through FullStory, Hotjar, and Posthog recordings to find insights or identify problems. With Decipher AI, you can skip the manual work—AI analyzes every session automatically and surfaces actionable insights instantly. Detect issues like errors or features not loading, answer product questions without manual instrumentation or SQL, and get real-time alerts about onboarding progress or key feature usage. Founders Rohan D. and Michael Rosenfield bring deep expertise from Google, having led multi-billion-dollar, multi-billion-user products at Google, including the YouTube homepage and Pixel. They created Decipher AI to make understanding user and product behavior effortless for everyone. Congrats to the team on the launch! 🚀 https://lnkd.in/dEGgPNqh