The most relevant direction for startups is the creation of specialized AI assistants for thematic niches. But how to choose a niche where they will be snapped up? It turns out there is a way to make such a choice. Otherwise, this startup would not have raised 3.6 million dollars in investments, even without opening public access to its AI assistant. And the essence of its approach is in the following.
The essence of the project
Basis has developed a specialized AI platform for accounting firms that handle financial and accounting management for other companies.
The Basis platform has not yet been launched publicly, and on the site, you can only add yourself to the waiting list. Nonetheless, the startup claims that they have already onboarded the first clients in a pilot mode. Despite such an early stage, Basis has just raised the first 3.6 million dollars in investments.
The first obvious functionality of the platform is a specialized chatbot with AI, which can answer specialized questions related to finance, accounting, and documentary support of companies’ operations.
A more interesting set of AI machine capabilities of the platform is that it can analyze and respond to questions regarding the data of each specific client whose accounting is managed by the firm.
The source materials for financial and accounting management are the list of client’s revenue and expenditure operations. To properly account for the operation, it is necessary to understand its essence. Basis can explain the meaning of each operation so that the firm’s employee can correctly attribute it to the respective revenue or expense item.
Basis can extract additional information about already conducted operations, as well as about the schedule of upcoming revenues and expenditures from the texts of contracts concluded by the client with suppliers and customers. This is another source of information about the essence of the accounted operations, and about what documentary evidence may additionally be required to confirm these operations.
Having all this information, Basis can answer general questions about the client’s data. For example, you can ask the platform why the client had such a sharp increase in marketing expenses in July-August?
And it will answer that this increase is due to the client transferring 8 thousand dollars to a certain company that deals with advertising — which most likely means that the client launched an advertising campaign in August. Therefore, these expenses can be attributed to a certain item, and also don’t forget to find from the client or request from the contractor the acts confirming the performance of work on creating and displaying advertising.
At the end of each financial period (month, quarter, year) all the client’s operations distributed by revenue and expense items are summarized into a single balance report, which the accounting firm should check again before signing it at the client’s and sending it to the tax authorities.
Basis can also perform this work. In this case, the AI machine will not only find possible inconsistencies, incompleteness, or contradictions — but also offer a list of actions that can fix this.
In the process of accounting, accounting firm employees regularly have to answer the client’s questions “why so,” “what does this mean,” and so on.
Basis can help with this too — automatically drafting responses to client requests, using not only the general knowledge base on finance and accounting but also the specific client data for whom it is composing the answer.
What’s interesting
Just a reminder, the main problem with universal AI machines like ChatGPT is that they periodically “hallucinate,” that is, invent nonexistent facts. Therefore, in their original form, they cannot be used for tasks where high accuracy of facts and conclusions is initially required.
This is precisely the reason for the sharp increase in the number of specialized AI platforms, sharpened for solving separate categories of tasks. These platforms are not only additionally trained on thematic data sets to understand terms, jargon, and other specifics of the subject area. Special databases with reliable facts are also connected to these platforms, which the AI machine of the platform uses to generate its answers. This technology is called “Retrieval Augmented Generation” (RAG).
Here are some examples of specialized platforms:
Rippey.AI — AI platform for logistics. Raised 4.8 million dollars in investments.
Mantle — AI platform on financial and investment issues for startup founders. Raised 7.68 million dollars.
Mindtrip — AI platform for travel planning. Raised 7 million dollars. A separate question on the topic of today’s startup — what is so interesting about the field of accounting reporting that a startup with a still unlaunched platform managed to raise 3.6 million dollars in investments?
One of the founding fathers of the USA Benjamin Franklin said: “In this world nothing can be said to be certain, except death and taxes” 😉 And the correct payment of taxes is what companies keep accounting for.
As a result, there are many accounting firms. Because many companies outsource this work to professionals. Only in the US, there are 88.5 thousand businesses offering a variety of services in the field of accounting.
And this is quite a market. Which has recently encountered problems.
More than 300 thousand people left the positions of auditors and accountants only in the US for 2020-2021.
Their places are taken by newcomers, who due to lack of experience cannot yet cope with this work at the proper level. Unless an AI machine like Basis helps them, of course 😉
By the way, newcomers are now reluctant to go into this profession, which was recently written about by The Wall Street Journal in the article “No one wants to be an accountant”.
Therefore, the “old-timers” have to cope with at least the same volume of work, but with fewer forces — in which again a specialized AI platform can help, significantly saving their time and effort.
Because otherwise, due to the lack of hands, accounting firms are even forced to start refusing clients.
Most often this has to be done to small companies — although for them every earned dollar matters. In 2022, accounting firms with revenue of less than 8 million dollars were forced to refuse 6.46% of clients.
But “the rich also cry”. Accounting companies with revenue of over 300 million dollars a year in 2022 were forced to refuse 2.4% of clients.
In other words, the demand for accounting services is high, the number of companies providing these services is large… but there is no one to work in them 🙁 And this is a problem! Which, apparently, without specialized AI platforms, allowing to perform larger volumes of accounting work with fewer forces, can no longer be solved.
Where to run
The general direction of movement is clear — the creation of specialized AI assistants for individual thematic niches. And you need to start doing this right now — while all the niches have not been snapped up by competitors 😉
However, which niche is best to go into? Where will such AI assistants be snapped up? Probably, where these assistants can help not just make good even better. And where they will help solve the exacerbated problem — that is, make good out of bad.
This means that an interesting criterion for choosing a niche for creating AI assistants is niches in which there is now an increase in the number of layoffs and a decrease in the inflow of new employees — as, for example, in the field of accounting.
And what other such losing attractiveness for employees niches can be? Then it is there that you need to stick AI assistants who would take on the tasks that people do not want to do.
Going into such niches may seem not sexy enough — but demand and money should definitely be there 😉