AI Training Data for Business LLMs: Train your LLMs with business emails

At Kirchner we have vast collections of industry specific business emails between manufacturers, brands, distributors and shops.

business emails are an incredible source of business wisdom and expertise which can be used to build large language models specialized in purchasing, sales, business management, and many other business fields.

Building an accounting LLM:

Lets assume you are building an accounting LLM that pre-answers all incoming emails for accounting firms, saving the accounting firm up to 50% of monthly man hours. All accountants do now is click on answer – the pre-written email pops up with all questions answered – the accountant can add small changes when necessary and then send out the email.

Using Kirchner’s accounting data:

Lets take for example the emails of an accounting firm, between for example 10 accountants of the firm and its clients over the range of 5 years. And lets assume each accountant writes 20 emails per work day. Thats 260 work days x 10 accountants x 20 emails = 52,000 sent emails, adding the received emails we reach a total of over 100,000 emails covering most questions and answers that come up in the interaction between accountants and their clients.

Building a shoe factory sales LLM:

Lets assume you are building a sales LLM that pre-answers all incoming emails for a specific industry or product. In this case let’s assume your LLM is pre-trained for shoe factories.

Your LLM in combination with Kirchner’s data enables to answer inquiries and follow up questions with one click.

Not only will this LLM save 50% of man hours of the shoe factories sales team, but also enable the factory to answer inquiries that where in the past too small but with automation now become viable revenue channels.

Using Kirchner’s accounting data:

Lets take for example the emails of an accounting firm, between for example 3 sales employees of the firm and its clients over the range of 5 years. And lets assume each sales employee writes 40 emails per work day. Thats 260 work days x 3 sales employees x 40 emails = 31,200 sent emails, adding the received emails we reach a total of over 62,400 emails covering most questions and answers that come up in the interaction between sales employees of a shoes manufacturer and shoe brands.