“Do you really understand all the terms and pitfalls of the contract that the investor slipped to you? And do you have the money to pay for legal assistance every time? It’s much easier and quicker to turn to a special AI assistant created by this startup. That’s already wonderful! Even more wonderful is that now there’s a recipe for creating similar startups in other fields.
THE ESSENCE OF THE PROJECT
In addition to developing and promoting their product, startup founders also have to deal with legal and financial issues related to owning their company. For example, what share can be given to investors for what money and on what terms, how to structure options for key employees, and how to properly formalize all of this.
However, founders usually aren’t big experts in the legal financial area, so in the process of solving these issues, they can make critical mistakes that will have negative consequences.
Mantle has created a platform that “came from the future to help founders solve ownership issues of their companies”.
The core of the platform is an AI assistant, which is well-versed in this range of problems, as it has been specially trained and educated on the necessary data set.
In its simplest version, it operates as a chatbot, to which you can ask questions and get answers – like “how will my share change after signing a convertible loan agreement?”. Questions can be general or include some specific conditions proposed by the investor to the founder.
Another way to use the platform is for the automatic execution of procedural tasks.
For example, the AI assistant can be entrusted with the task of allocating options to key employees on the terms set by the founder. Then the assistant will prepare the necessary set of documents, send them for approval to the board of directors, wait for their electronic signatures, and then send the contracts for signature to the specified key employees – after which it will report the results to the founder.
However, the question arises – how can the founder properly set the terms for allocating these options? But Mantle also has an answer to this question 😉
For this, a special module for document analysis and scenario development has been designed on the platform.
You can upload, for example, a term sheet (a preliminary document with the terms of a possible deal) received from an investor. Even if it contains a bunch of complex terms and conditions — the AI assistant will figure them out, provide a brief summary in the form of specific numbers of share changes after the contract is signed, and can also answer the founder’s questions like “what will happen with this?” or “what will happen if this occurs?”.
The platform’s functionality is not limited to the above examples. It will help founders navigate a wider range of questions related to fundraising, share distribution, financial forecasting, and even hiring plans. All documents, scenarios, and plans will be stored in one place, eliminating the need to search through email boxes and gather data from tables scattered across different folders each time.
Moreover, on the platform, you can set up standard procedures for document distribution and approval with the necessary participants in the process — then launch any of the procedures according to a schedule or at the push of a button and monitor its progress on a special dashboard.
At each stage of startup development, founders are concerned with very different questions and problems, but the AI assistant of the platform is well-versed in all of them. Therefore, Mantle can accompany founders throughout the journey — from initial investments to going public or fully selling the company.
Mantle was founded last year and has now raised its first investment of 10.5 million Canadian dollars (7.68 million US dollars).
What’s interesting
Mantle is at the intersection of several trends.
The first trend is the creation of AI assistants for solving specific types of tasks.
The problem is that universal chatbots with AI like ChatGPT can “hallucinate” — inventing non-existent facts, referring to non-existent laws, describing impossible actions, and so on.
It’s somewhat forgivable for a universal AI assistant, but specialized products cannot be forgiven — because they are bought to solve specific tasks, and hallucinations are unacceptable here.
Therefore, such specialized AI assistants usually work according to a slightly different algorithm of “Retrieval Augmented Generation” (RAG). In this case, special databases containing factually accurate information are used for providing answers, and the information extracted from these databases is automatically pulled to prepare the answer, ensuring its high accuracy.
This means that for each specific area of application of AI assistants, it’s necessary to collect such databases, keep them up-to-date, and link them with the current versions of the universal AI engines being used. This represents a separate effort – and an opportunity for creating individual startups focusing on this.
Among such startups you could offhand remember:
- Mindtrip – AI assistant for travel planning. Even before launching their platform, they managed to raise 7 million dollars in investment.
- Curipod- AI assistant helping school teachers prepare for lessons. Raised 4.8 million dollars.
- Sizzl – AI assistant helping schoolchildren and students with homework solutions. Raised 7.5 million dollars.
- Rippey.AI – AI platform for solving specialized tasks in the field of logistics. Raised 4.8 million dollars.
The second trend is the active penetration of AI assistants into the legal field, which has become one of the growth drivers of the legal tech market.
The problem is that contracts are complex documents, abounding with special phrases and terms, which are difficult for an unprepared person to understand.
Creating such documents is not easy even for specially trained people. Because everything needs to be very clearly formulated with a bunch of subtleties and details, and the price of the slightest mistake can be very high both for the contract participants and for the lawyers composing the contract.
AI can help a lot with this, as it can instantly analyze very complex documents, extract all the details from it, check them for consistency with each other and with past precedents – thus acting as a very smart assistant capable of instantly delivering the results of a comprehensive analysis.
As an example, the startup The Contract Network can be mentioned, which created such an AI platform and raised 8 million dollars in investment for it.
The third trend is that AI is starting to become an assistant to founders who do not have the opportunity to hire large teams to launch their products.
Today’s Mantle helps save on hiring financiers and lawyers. Another startup created the Gobi platform, which can replace a product manager to help the founder develop a product creation plan. After which some tasks are delegated to AI for resolution, and for the remaining tasks, performers from the existing marketplace on the platform are hired. This platform has not even been launched yet, but it has already received the first 400 thousand dollars in investment.
Where to run
Helping startupers is, of course, a noble cause 😉 But the startup market is not that big. Moreover, most startups, unfortunately, die at quite early stages, which does not allow earning on them for a long enough time 🙁
Therefore, the main direction of movement is still the creation of specialized AI assistants. Especially in those areas where there are a lot of details and subtleties, special concepts, complex documents, and/or where the cost of low accuracy of answers from universal AI assistants is too high.
In legal or logistics themes, all these conditions are met. And in the theme of lesson preparation, there are no major difficulties with school material, but the level of accuracy of AI assistant advice needs to be extremely high.
The general scheme of building such specialized AI assistants is already clear. So the main thing is to find a suitable area for their application, so that the market is already big or at least growing.
For which area or for solving what types of tasks would it be worth creating such an AI assistant now? While no one else has captured this topic 😉
About the company
Mantle
Website: withmantle.com
Latest round: $7.68M, 13.10.2023
Total investments: $7.68M, rounds: 1