A Montreal startup wants to give accountants more time to advise clients by allowing AI to process documents.
Nima Jalalvandi of Montreal says every time tax season rolls around, it brings him nothing but stress. As someone who paid taxes for his family and friends, Jalalvandi said it was like “a spring cleaning, without any benefits.”
Jalalvandi and his co-founders responded by creating Ready plan aheadan artificial intelligence (AI) platform that aims to automate routine paperwork for accountants and allow them to spend more time providing clients with tax and financial planning advice.
“We want to increase satisfaction so accountants feel they can do meaningful work.”
Nima Jalalvandi, Ready Plan Go
Ready Plan Go calls its software an “AI-powered back office” for tax preparers. It is designed to speed up document collection, data extraction from digital documents, and tax form entry software.
“We want to increase satisfaction so that accountants feel like they can do meaningful work by creating a back-office infrastructure layer for tax preparation,” Jalalvandi told BetaKit.
Having completed the Station FinTech Montréal and Next AI accelerator programs, Ready Plan Go received two Startupfest awards worth $175,000, including a 2SLGBTQIA+ award and a FinTech grant. The early-stage company has now increased that to a $750,000 pre-seed round, including a check from Mistral Venture Partners and 15 angel investors from Montreal's startup ecosystem. They include Stay22 CEO Andrew Lochhead and Firat Gezen, former president of General Dynamics Canada. Part of the money came from the sale of Jalalvandi's own house.
Ready Plan Go's business model depends in part on acquiring the books of tax firms whose accountants are leaving the business or retiring. The company primarily targets U.S. firms because they have a larger customer base, and it recently landed its first major client in Pennsylvania.
CONNECTED: Accounting startup Ceedar uses artificial intelligence to turn clients into investors
Jalalvandi said the money will be used to buy more businesses and work on product development. The four-person team consists of two founders with business backgrounds and two chartered professional accountants.
Although the AI platform is intended to replace some of the work of accountants, Jalalvandi argues that it only reduces the workload of a dwindling workforce. The accounting industry is on the cusp of a major transformation, with more than 80 percent of licensed professionals expected to retire within the next decade and fewer young people joining the profession. An April 2024 Survey A study by Robert Half found that 90 percent of hiring managers in finance and accounting are already facing challenges due to labor shortages.
Jalalvandi sees the accountant shortage as an opportunity for companies to integrate Ready Plan Go technology. The online platform features AI agents that perform multiple steps of the tax preparation process, as well as connecting to external filing platforms such as Drake and ProSeries.
“There will be a person there who will make sure that the context is always in place,” Jalalvandi added.
Some research into the use of AI in accounting has shown that it is useful for speeding up tedious “front work” such as classifying transactions and collecting customer information. According to Stanford University Poll Of 277 accountants at 79 U.S. firms, accountants using AI served more clients, completed monthly reports a week earlier, and spent less time processing back-office tasks. However, the same study found that 62 percent of accountants are concerned about errors created by AI, and 43 percent are concerned about data security when using AI.
Jalalvandi explained that clients of the firms he works with will have to consent to sharing their data with these AI models. He added that the company is working to achieve SOC 2 certification, a widely used security framework.
Ready Plan Go also uses a combination of large language models (LLMs) to ensure that tasks are completed with maximum accuracy. Jalalvandi gave an example of extracting text from PDF documents: If several LLMs do not achieve 100 percent accuracy, the system sends a request to a human to review the documents. “Ninety-nine percent is not enough,” he said.
Image courtesy of Ready Plan Go.
