Lutra AI

Introducing Lutra, an AI Agent that generates and runs workflows for you. Lutra first converses with you to understand your goals, and then generates and runs code that integrates with the apps you use. Lutra has crafted an environment for AI generated code to run in a production-ready setting. By taking a code-first approach, we leverage the best practices from software engineering for: 1. Security 2. Maintenance 3. Governance We're excited to announce $3.8M in funding from Coatue, Hustle Fund, Maven and WVV Capital, together with Angel Investors that include Andrej Karpathy, Jeff Dean, Scott Belsky, Quoc V. Le, Tom Willerer, Brian Shultz, Vijay Karunamurthy, John Ciancutti, Irwan Bello, Keith Peiris, Kah Seng Tay, & more! Sign up today at https://lutra.ai Follow us for on X/Twitter at https://lnkd.in/gBx4Tpzn for updates Join our discord at https://lnkd.in/gZZfJqSy

We are introducing Kits on Lutra AI! Kits are collections of flows that work together to automate an entire process. Our first kit is a Lead Gen Kit that helps you find accounts, look up decision makers, their contact information, and assist with the reach out. In this first Kit, you get workflows that: - Use Job Postings to help you identify relevant accounts. - Find decision makers using role based search. - Find email addresses for the decision makers. - Draft custom reach outs, integrated with your email service. What other Kits would you like to see? Let us know -- we'd love to work with you to make them. More at https://lnkd.in/gsncvD2x

We were onboarding Tom Goodwin today who works film industry; they wanted to extract information from their invoices and get that into a database. These invoices contained vendor information, lines for each item (e.g., led lights) purchased, prices, quantity. They wanted that data extracted for a whole folder of PDFs and placed into a spreadsheet. We were able to automate this easily: pointed Lutra to the Google Drive folder and gave it instructions to extract the line items, and put all the data extracted into a new spreadsheet. If this sounds like a problem you run into a lot, comment/let me know!

We're excited to share how customers like Fifty Years leverage Lutra AI to enable their teams to have an outsized impact. A few weeks ago we met Kia Fathi and Drew Moxon and introduced Lutra AI to them. Since then, Lutra has become an integral part of their workflow for their 5050 program. They have used it to process large datasets, scrape thousands of websites for talented individuals, and enrich them with relevant information (LinkedIns, Career History, Contact Info), and classify and sort them. The natural language interface for defining workflows has been incredibly helpful, alongside the pre-built integrations including Airtable, Google Sheets, People Data, LLM integrations, and advanced scraping capabilities. Read this case study to learn why 50Y chose Lutra after evaluating Windmill, Retool, Gumloop, Browserbase, Zapier, Autotab, Intuned, and rolling their own solutions! https://lnkd.in/djPA5iUD

Have you had to send dozens or hundreds of emails ?? after an event or to a group of people, customizing them differently for each? We've been onboarding users in groups, and here's my workflow for creating a batch of these emails. With AI tools like Lutra AI ??, it turns out that creating personalized drafts can be as easy as telling the AI the table you want it to use, and the template for the email. This video shows how you can set all that up in under 2 mins. ?? Taking another step towards freeing us from the mundane and repetitive tasks!

Generative UX is the next frontier for AI applications. Most AI applications are based in chat today -- while chat is flexible, it has many issues. Generative UX that creates custom experiences on the fly could solve a lot of these problems. Two big issues we encountered was the (a) lack of affordance: it can be hard to understand what a chat-based system can and cannot do -- and in the case of AI, it's hard to "point and give feedback" to an AI system, (b) inefficiency: for frequent actions, being able to directly tap/click on a UI element is much simpler than asking chat to accomplish it; overtime, too much chat can lead to fatigue. The idea of "prompt engineering" is a temporary stop-gap for dealing with these issues. A lot of UX in today's business applications tend to be reflection of the underlying data representation, often resulting in endless menus and options. But, what if the UX was instead on-the-fly tailored to the user's desired outcome? Can we automatically determine the appropriate controls to show, and also have affordances to let them dig deeper. As we get into the world where each of us can build our own custom software, generative UX will be a big piece of the puzzle to figure out! [1] https://lnkd.in/gQwvw5n2 [2] https://lnkd.in/g_Vb_Hve [3] https://lnkd.in/gmw_mCDk

Why are AI Agents exciting? Even after a year of AI hype ??, integrating AI into our day-to-day workflows remains challenging. For example, it is not straight-forward to get an AI to repeat a task over a large spreadsheet, or to look at every email - extract relevant data and send it to Airtable, or to process a folder of PDFs files to extract key terms into a spreadsheet. We know if we copied and pasted the data into ChatGPT or Perplexity, they can do intelligent work with our documents. But the manual copy-paste steps get in the way of the promise of real productivity gains from AI. You need to code (or get AI to code) the integrations -- or use old-school workflow integration solutions that are visual-based (e.g., Zapier). However, these visual programming interfaces are often hard to use and quickly become a messy spaghetti of connections. I think this is why many people are excited with AI Agents. They bring the promise of using language to direct the computer to do all these work, with the use of AI to process data. The idea that you can tell a computer what to do and it'll get it done. We're building headed towards a future where each of us can use AI to create our own custom AI tools that work the way we want them to, and are fully integrated into the apps we use. Exciting times! ?? #ai #aiagents

Useful AI Agents are here! ?? One highly manual task in developing a potential client list is to go through each potential company/contact and do research on their industry, size, funding, team, sentiment, and more. Lutra can do this automatically for you. The data is often all over the Internet, and it can take quite a long time to find all of them, put it together, and then make a determination on fit. ?? ?? Here's where AI Agents like Lutra can help! Simply tell it what information is relevant to you and Lutra can search the web to find the right information for you. Here is an actual prompt that you could give Lutra ?? "For each company domain, go to the website and: Extract the company name and company social links. Summarize what the company does. Update the columns accordingly. Fill out each row as you process data before moving on to the next row.” ? This works with Google sheets, Airtable, HubSpot, and can even be a Webhook. Learn more on our blog at https://lnkd.in/gBkEeTkY where we go into details on how this works. If this is something you'd like to try, sign up today! ??

AI Agent systems need to make this distinction in order to be effective: 1. Using AI models for figuring out what tools to use, and how to use them. 2. Using AI models as smart tools for data processing. One of them requires high level reasoning (hard task!), and the other might be much simpler (e.g., summarize some text). Depending on which mode you're operating in, you will want different models - large frontier models for good reasoning; and small specialized models when using them as tools. If your AI Agent platform is not making this distinction, then you're likely operating in an inefficient approach: slow and high costs.

It's always a highlight for me when we are able to help solve an actual problem within the first onboarding call! I was showing Clarence Chio how Lutra AI works today. In just a few minutes, we were able to grab a list of speakers for him, use AI to create one-liner summaries of each speaker bios, and also find their LinkedIn profile links. If you're heading to TPRM summit in Atlanta in June, this list might be handy for you too! Original website: https://lnkd.in/gDYJHHMY List extracted and AI enriched: https://lnkd.in/gy-56KiG Here's how it worked: https://lnkd.in/g62DNHS4