Working on this w/ tryZeroShot.com — it’s agent session monitoring with Signal extractions we spent 6 years developing, identified where and when you can build a skill that speeds up sessions by 3-7x, reduces tokens by up to 90%, and can tell coding agents which models perform the best for what task. Has a 100% local version that’s free.

a b2b chat agent that does stuff even Claude Code can’t. we built something different. 3 things our new agentic chat does that i haven't seen anywhere else: 1. quantitative analysis, not vibes. when you ask "summarize feature requests from the last 30 days," you get the same answer every time. not a semantic search that gives you different results depending on the mood of the embeddings. real numbers. real signals. click into any one and see the actual customer conversations behind it. 2. MCP integrations across your entire stack. ask one question that pulls from PostHog, Linear, and your customer signals simultaneously. "what are our top customer issues, ranked by usage, with associated tickets?" — one prompt, three tools, one answer. no tab-switching. no copy-pasting between dashboards. 3. your full company context in one place. plug in your knowledge base — pricing docs, roadmaps, product documentation — and combine it with customer signals, internal conversations, and every integration you've connected. it's not a chatbot with a template. it's fully agentic. no deterministic workflows. any question, any answer. the part that I love: methodology transparency. every response shows you exactly how it analyzed the data, so you can refine your questions and understand what's happening under the hood. this is one of the best agents i've seen in b2b. and it ships with unlimited access for everyone. p.s. @BuildBetterAI agentic chat is live now — 2 min demo below

BuildBetter’s MCP is Live! 🎉 What can you do with it? • Claude: “Review customer issues from the past 6 months and cross-reference Linear tickets. Summarize which issues we’ve resolved.” • Cursor: “Get my most recent call with Adam, and write a scoping doc for Cursor to implement the feature we discussed.” What is MCP and why should you care? MCP (Model Context Protocol) lets your AI seamlessly query your BuildBetter data (customer and team calls, surveys, insights, documentation, chat messages, tickets, etc.)—no need to figure out complex APIs. Our local MCP server can be installed and running in minutes, compatible with any MCP client, such as @AnthropicAI's Claude, @n8n_io, @Cursor_AI, @windsurf_aI, @SlackHQ and dozens more. Mix & Match to Automate Workflows: BuildBetter’s MCP seamlessly integrates with hundreds of other MCP servers. Instantly build powerful workflows connecting BuildBetter data with nearly every tool in your stack. Here’s a personal favorite of mine using Claude + @Linear + @BuildBetterAI:

OpenAI launched Deep Research, but it can’t touch your internal data. We fixed that. Meet Deep Analysis—turning a year of customer feedback into actionable insights, prioritized by ICP, aligned with your roadmap, and backed by real quotes. 100+ integrations. Your internal data. Instant clarity. Early access is free: BuildBetter.ai

Company swag is a waste of capital. So why are we launching (Build)Better Wear apparel? Because we’re doing it differently. Something you’d actually buy. Something you’d actually wear. Something you’d actually show off. High quality—no generic Canvas/Levels tees. Our first collection is called: Raise. Apparel we’d wear ourselves, printed and embroidered with sentiments that matter. Like our VR ➔ Crypto ➔ AI shirt—for founders chasing the next big trend. Our "pmf." hat—to declare you’ve found Product-Market Fit. And our “Build in Private, Brag in Public” sweater... because... well... that's what usually happens. If we’re making clothing, it has to be top-notch. It has to speak to you. It has to feel different. If you’re a customer, schedule a demo call and get free clothing—just mention you saw our apparel. We’re excited to hear your thoughts. We love them. We think you will too. Checkout the shop (limited launch): shop.buildbetter.ai Book a demo to get this apparel free: buildbetter.ai/onboarding

Over 10% of our paid customers were previous users at their former company. We’re often the first tool they sign up for when they start. Wild.

We just ran an eval test @ BuildBetter.ai for the new OpenAI o1-preview over our Signal extractions (customer issues, complaints, features, etc.) for Customer + Team calls. One example of extractions from a customer call we had: gpt-4o found 11 extractions. gpt-o1 found 9 extractions. We then had Claude evaluate 4o's and o1's extraction capabilities (why not). We also human reviewed both extractions and comparing them to the transcript, we found, o1 was overall better. Here's what we saw: - Accuracy and Detail: o1 provides more accurate and detailed summaries of the conversation. For example, it captures specific interest in extracting answers to five key questions from sales calls, which is a crucial point in the conversation. - Emotional Context: o1 does a better job of capturing the emotional nuances of the conversation. It noted frustration with X competitor and their excitement about potential features. - Relevant Business Information: o1 focuses more on the important business aspects of the conversation, such as their needs, pain points, and decision-making process. - Chronological Accuracy: o1 follows the flow of the conversation more closely, which makes it easier to understand the context and progression of the discussion. - Specificity: o1 provides more specific details about requirements and responses, such as the mention of bi-weekly reporting and the research assistant feature. While 4o does capture some important points, it misses some key details and nuances that o1 includes. 4o also contains some irrelevant information (like the discussion about living in X city) that doesn't add much value to understanding the business conversation. That being said, 4o does have some strengths: - It captures more technical details about BuildBetter's features. - It provides more information about competitor comparisons. ---- We're going to run more evaluations this week. But, if you want early access to o1 for your team that is HIPAA/GDPR compliant, and use o1 to chat with your customer data (like calls), write documentation, or get Signal extractions, sign up for free, and comment below and we can turn it on for you! Sign up: BuildBetter.ai Excited for this year!

Every AI startup: "Look how fast we can do X!" Me: "But can you do it well?" Why are we using AI to flood the world with mediocrity and spam instead of elevating quality? I dove deep into the quantity trap of AI and why the focus on output over outcomes is do harmful. From sales emails to content creation, we're getting AI all wrong. Since building on LLMs since 2020, we've thought a lot about this topic @BuildBetterAI . We hope this can help guide new AI founders on where to focus.

We're hiring. Sam Altman said, “Burnout actually comes from failing and things not working. Momentum is really energizing the lack of momentum is super draining." I always forget about holidays, so I pinged our team, offering to take their onboarding calls yesterday. It’s been 4 years for me, and I still get excited to work every day. This screenshot was from someone on our team. I hope you all get to work for a company where you feel like this on Mondays. If you think BuildBetter.ai would be that company for you. We're hiring, and we'd love to have you.

The stuff we have been cooking @ BuildBetter.ai is wild. Massive companies are pre-paying and setting up their accounts to get early access—this release is the start of what is to come. The biggest thing is: Intercom Integration is here and it's unlike, anything you've seen or tried before. When you add Intercom to BuildBetter, we process every ticket by type, topic, sentiment, emotion, and impact, allowing you to analyze trends and find specific quotes easily. You can even chat with every support ticket in combination with your internal data. It's accuracy is unlike any other platform. But because BuildBetter doesn't just have your external customer data, but also your internal calls and chats... for the first time, you can answer questions like: "What are our customers complaining about this week in Intercom that our team isn't working on?" If you've ever tried to answer that question, you know the weeks worth of work. We've also added 3 of our largest requested quality-of-life improvements: 1) Individual call sharing: You can now keep a call private and share it with one person. 2) Bot image options: Now you can customize what the bot looks like for your team/org calls (small, but big). 3) Join Workspace URL: Share your organization invite link instead of needing to invite others one-by-one. Plus, we still have domain join verification steps, so it's not just the fastest but the safest way to get your team onboarded!