Blogs, Updates & Announcements

We’ve shipped evaluations, a product redesign, embedded analytics via API, user retention analytics, alerts for areas of poor performance, a new ingestion model, manual ratings, and many smaller improvements.

What started as a tool for the Context.ai engineering team became an open-source utility we're now excited to be sharing.

We're proud to announce this major vote of confidence in Context.ai's commitment to user privacy and security.

What are the security risks of deploying LLMs to production and what can you do to stay prepared?

We've overhauled the Topics page, added pages for you to track individual users, and a whole lot more this month.

Don't let your LLM become a liability. Proactively monitor for toxic behavior to keep your AI respectful.

Intelligence isn't everything. Sometimes, LLMs are just too slow for good user experiences.

LLMs are a great way to cut costs. But they're also an easy way to blow your budget.

LLMs are notoriously fickle and prone to making up information.

Retrieval-augmented generation bridges the gap between LLMs and your private knowledge base. In this post, we’ll introduce RAG and discuss how to know if it’s the right strategy for your AI app.

Prompt engineering isn’t just for engineers. In this piece, we’ll apply a product lens to prompt engineering. We’ll review fundamental prompting techniques and the importance of iterating on prompts according to user feedback.

AI isn’t just a frontier for engineers. It also presents complex challenges for product leaders. In this post, we’ll walk through the AI product development cycle and show you how to improve user experiences even when they’re highly personalized and rarely identical.

Check out the latest features including the new Workspace and Users pages, expanded analytics capabilities, and PII filtering for protecting sensitive data.

Startups need to be 10x better than the competition to disrupt a market. With generative AI being a 10x technology in many ways, let’s explore how AI-native startups are using it to create outsized value.

At Context.ai we recently completed a $3.5M first fundraise in a heavily oversubscribed round that you can read about here. I’ve since been asked for advice by many founders about how to approach an early stage fundraising process

Today we’re announcing that Context.ai has raised $3.5 million in funding co-led by GV (Google Ventures) and Tomasz Tunguz at Theory Ventures. Context.ai provides product analytics for LLM-powered applications, helping businesses build better AI products.

Working in an office is a huge competitive advantage. The default for a startup in 2023 seems to be remote-first. I think this is a mistake.

Today we’re sharing an overview of how ThoughtSpot builds products powered by Large Language Models, and how the team uses product analytics to guide development. This blogpost is the result of an interview with Jasmeet Jaggi, VP Engineering at ThoughtSpot, who leads a team building ThoughtSpot Sage, a natural language search driven analytics product.

You wouldn’t launch a product or marketing site without analytics in place to track their performance with real users. Why not? Because you need to understand how real people are using your products and how well they meet user needs.

Before I started Context, I often wondered what starting a startup would be like, and how it would differ from my previous job as a Product Manager at Google. This is what a typical week looks like, and how that differs from my expectations.

As an AI startup founder in London, one of the most common questions I hear is “when are you moving to San Francisco?”. California is clearly a great place to start a startup, but in 2023 I think London is even better. Building a successful startup requires many components: talented employees, access to capital and experienced investors, a startup ecosystem, and a market of potential customers. London scores exceptionally on these dimensions.

Aggregators of consumer attention earn enormous value on the internet. These businesses connect large and disconnected groups of buyers with sellers, controlling consumer attention, and charging sellers for the privilege of reaching the consumers. ChatGPT and ChatGPT plugins have the potential to disrupt these aggregators by changing the behavior of consumers. Yet today OpenAI is leaving value on the table.

Yesterday, researchers at Stanford and Berkeley published a fascinating paper showing that LLM performance drifts over time. I don’t think anyone would be surprised that the OpenAI GPT4 service performs differently to GPT3.5. But even within GPT4 or GPT3.5, performance shifts dramatically, including significant worsening.

Today we’re announcing a Langchain integration for Context. This integration allows builders of Langchain chat products to receive user analytics with a one line plugin.

I quit my job at Google to work on the future of AI chat products. I deeply believe in their potential to deliver amazing new experiences, but obviously I don’t expect anyone to take my word for it. I’ve talked to approx 100 developers of LLM-powered chat products, and I’ve learned interesting lessons about the future of the space. 

Chat interfaces powered by AI can deliver amazing user experiences. Companies from law to medicine and e-commerce to entertainment are building these products at breakneck pace. But while building an impressive demo has never been easier, developing a production-ready product is tough.