AI tool Jenesys lands $1.1m to reimagine the role of accountantsby
Tech startup Jenesys AI has secured a pre-seed funding round of $1.1m to elevate the role of bookkeepers and accountants with an artificial intelligence co-pilot built to perform a variety of back-office accounting functions.
Powered by Jenesys’s proprietary artificial intelligence (AI) technology, the firm has launched co-pilot tool ‘Jack’ with the ambitious goal of improving transactional accounting and shifting accountants from ‘operators’ to ‘validators’.
According to a release from Jenesys, once Jack is up and running it will perform a range of back-office tasks, including bookkeeping reconciliation, fraud and compliance checks, approving and making domestic or global payments, and updating budgets, effectively operating as everything from bookkeeper to internal auditor and ensuring compliance to accounting policy and international financial regulations.
While Jack is currently still in development, according to CEO Nicolai Thomson the first product off the Jenesys AI production line will be a bookkeeper you can hire for £21.60 an hour that will complete 120 reconciliations in that time, including line items, VAT and general ledger codes ready for payment.
Using Teams, Slack or email (with a WhatsApp integration coming in early 2024), clients or employees will be able to submit invoices or expenses into the system by sending a picture or PDF, without having to push a specialised finance app and train clients on its use.
The tool will offer approval and budget functionality if required, where Jack will check against existing company policies, and even auto-approve and pay invoices or expense claims.
Accounting Language Model
Jenesys has taken the principle of the Large Language Model, the technology behind generative AI tools like ChatGPT and Google Bard, and fine-tuned it to produce its own proprietary ‘Accounting Language Model (ALM)’ that powers the Jack co-pilot.
According to CEO Nicolai Thomson, Jack has had more than 4,500 hours of training on 220,000 practice invoices and receipts. A portion of these have come from the company’s first acquisition, Danish tax tech company VAT Key, which has also provided pan-EU documentation for training. The acquisition of VAT Key also offers Jenesys the ability to leapfrog payment compliance issues and execute global payments at the same price as Wise and Revolut.
“Right now you can upload invoices, receipts or bank statements to ChatGPT and ask it to extract data,” Thomson told AccountingWEB. “But if you ask it to pick an account code that Xero uses for reconciliations it’s unlikely to be able to do this. If you ask it to calculate VAT on a line-item basis, ChatGPT needs to figure all this out itself, find VAT tables from HMRC etc. At best you’ll hit character limits or time out, at worst you’re likely to find inaccuracies.
“Jack is already trained to do these things – these behaviours are already in its long-term memory,” he continued. “We’ve built a finely tuned specialist and it’s not trained for anything other than accounting regulations. There are between 20 to 80 data points on each invoice or receipt, and we’ve used our team to annotate every single one as a foundation for Jack’s fine-tuning.
“The performance of Jack’s ALM is exceptionally high from a data extraction and categorisation perspective. Correcting VAT and category code extraction errors is often where accountants spend their time. Our aim is for Jack to take over most data entry and operational tasks, enabling accountants to focus on validation and analytical work.”
Jenesys is currently in development. To register an interest, accounting professionals can visit the company’s practice page and add their email address to join a waiting list. You can watch a demonstration of Jack’s AI ALM data extraction and AI approval capabilities here.
The Jenesys journey
Founded in November 2022 by enterprise and financial management specialist Nicolai Thomson and computer scientist Tosin Dairo, Jenesys had a presence at accounting trade shows in 2023 and established partnerships with several accounting firms in the UK to conduct preliminary testing of its Jack co-pilot.
The $1.1m pre-seed funding round was led by veteran AI investors Nick Slater and John Spindler at London’s Twin Path Ventures, with participation from Fuel Ventures, Antler, Hatcher+ in Singapore, and angel investors from the US and UK.
According to the Jenesys release, the new funding will be spent predominantly on engineering, customer success and expanding global payment compliance.