Skip to content

Dusty boosts team productivity with state of the art language models


Mud: Enhance crew productiveness with AI-powered choices

Mud, a recent AI startup based in France, goals to spice up group productiveness by breaking down inside silos, leveraging vital information, and offering customizable inside apps. The corporate makes use of Large Language Fashions (LLM) on inside firm data to empower workforce members with new capabilities.

Co-founders with a robust background

Based mostly totally on Gabriel Hubert and Stanislas Polu, Mud advantages from the extreme expertise and relationship between the duo, spanning over a decade. Their earlier startup, Totems, was acquired by Stripe in 2015. After the acquisition, each founders spent a number of years at Stripe earlier than branching out on their very own.

Stanislas Polu joined OpenAI and spent three years enhancing the reasoning abilities of the LLMs. In the meantime, Gabriel Hubert has taken on the position of Product Supervisor at Alan. Their paths converged as soon as extra after they joined forces to seek out Mud. Particularly, Mud differentiates itself by specializing in targeted LLMs developed by firms similar to OpenAI, Cohere and AI21, fairly than creating enormous new language fads.

Streamlining internal understanding with clay

In its early days, Mud targeted on constructing a platform for designing and implementing huge language dummy capabilities. The agency subsequently shifted its focus to centralizing and indexing insider data, making it accessible to LLMs. To perform this, Mud will depend on connectors that repeatedly pull information from in style platforms like Notion, Slack, Github, and Google Drive. The information is then listed, permitting for semantic search queries. When prospects have interaction with Mud-based capabilities, the platform retrieves the related insider data, makes use of it as context for an LLM, and delivers the precise options.

Mud transcends the capabilities of normal inside search software program by not solely producing search outcomes. It excels at extracting information from quite a few information sources and presenting choices in a medium that’s terribly helpful to the actual individual. Mud can perform as an inside ChatGPT, permitting for seamless communication and likewise serving as a basis for creating new inside units.

Gabriel Hubert expresses his notion of the transformative vitality of pure language interfaces, stating: We’re delighted that the pure language interface disrupts the software program program. 5 years from now, it is going to most likely be disappointing for those who nonetheless need to go clicking Edit, Settings, Preferences, to find out that your software program program ought to behave in a different way.

Varied capabilities of the software program program

Mud is collaborating with different designers to provide you with some methods for distributing and packaging their platform. Stanislas Polu highlights his creativity and clairvoyance, saying: We want there could possibly be quite a lot of completely totally different merchandise that could possibly be created on this home of company data, information people and fashions that could possibly be used to assist them.

Whereas Mud is but to be in its early phases, it faces a urgent downside. The LLM mix poses many obstacles, together with data retention and hallucinations. Nonetheless, because the LLM expertise progresses, these elements could steadily lower. Alternatively, Mud can develop their very own LLM to make sure the confidentiality of information.

Financing and future prospects

Mud not way back raised $5.5 million (€5 million) in an preliminary funding spherical led by Sequoia. Different contributors embrace XYZ, GG1, Seedcamp, Be A A part of, Motier Ventures, Tiny Supercomputer, AI Grant, and notable enterprise angels similar to Olivier Pomel of Datadog, Julien Codorniou, Julien Chaumond of Hugging Face, Mathilde Colin of Entrance, Charles Gorintin and Jean -Charles . Alan’s Samuelian-Werve, Pigment’s Eléonore Crespo and Romain Niccoli, BlaBlaCar’s Nicolas Brusson, Airtable’s Howie Liu, PhotoRoom’s Mathieu Rouiff, Igor Babuschkin and Irwan Bello.

In brief, Mud is betting that LLMs will revolutionize one of the simplest ways firms work. Considerably, the platform caters to firms that prioritize radical transparency, written communication, and autonomy. By leveraging LLMs, Mud unlocks untapped potential for the IT workforce, offering a aggressive edge to firms that embrace these values.

Repeatedly requested questions

1. What’s mud?

Mud is an AI startup that focuses on enhancing group productiveness by leveraging giant language fashions (LLMs) and inside enterprise insights. Its goal is to interrupt inside silos, simplify information sharing, and supply instruments for constructing customized inside capabilities.

2. How does Mud use the LLMs?

Mud makes use of LLMs developed by firms similar to OpenAI, Cohere and AI21 to empower group members with new abilities. Use connectors to tug information from platforms like Notion, Slack, Github, and Google Drive. This data is then enumerated and used for semantic search queries, offering sure right options to patrons’ queries.

3. What are the models of Mud other than totally different AI startups?

Mud differentiates itself by specializing in constructing options on high of present LLMs fairly than creating enormous new language fashions. Its goal is to streamline insider data and supply a clear and pure language interface to enhance communication and productiveness.

4. Is Mud merely a analysis software program program?

No, Mud outperforms typical search instruments. It not solely fetches search outcomes but additionally fetches information from quite a few information sources and presents them in a simple to make use of format. Additionally, it may be used as an concept to create new inside units inside an organization.

5. How does Mud profit information addicts?

Leveraging the framework of the LLMs, Mud unlocks untapped potential for the IT workforce. It allows seamless communication, tailors the software program program to folks’s wants, and delivers worthwhile insights from inside data sources, enhancing productiveness and effectiveness.


Mud’s stylish approach for enhancing group productiveness by leveraging LLMs and optimizing insider data has the potential to revolutionize the way in which firms work. By offering customizable inside capabilities and enhancing communication by way of pure language interfaces, Mud empowers your information workforce and unlocks untapped potential. Together with its present spherical funding and dedication to exploring quite a few utility potentials, Mud is poised to make a big impression on one of the simplest ways ahead for office productiveness.


To entry extra data, kindly discuss with the next link