Large companies have invested heavily in recent years in software designed to automate routine back-office tasks, many of which involve cutting and pasting data from one software application to another or Includes using drop-down menus to populate database fields.
Known as “robotic process automation” or RPA, this type of software “robots” are not AI. Some souped-up versions of Excel macros do little more than record mouse movements and keyboard strokes. Other software uses “if-then” rules to help complete workflows.
And yet businesses are estimated to spend more than $6 billion annually on RPA software, a figure growing at a double-digit percentage clip, according to technology analytics firm Forrester Research. UiPath is one of the leading players in the RPA field, valued at $13.5 billion. Appian, Blue Prism, and IBM also offer RPA solutions.
Now a small startup based in Munich, Germany, called Interloom, thinks it can disrupt this entire market by reinventing process automation on top of a new wave of big language models and generative AI assistants. Is. Air Street Capital, a London-based venture capital boutique that specializes in early-stage AI investments, has given the company $3 million in seed funding to hone its vision.
Fabian Jakobi, Co-Founder and Managing Director of Interloom, is a serial entrepreneur. In 2021, he sold his last company, Boxplot, to New York-based AI software company Hyperscience, which specializes in data extraction from unstructured documents. Jacoby believes that similar AI methods could be used in the future to extract information from video recordings, call logs, notes, etc., enabling AI software to learn what professionals are doing. How do people work? Then AI agents, based on the same basic AI methods that power today’s big language models, can be used to automate many parts of these tasks.
This can allow much more high-value tasks to be automated than can be solved with today’s RPA, which only works well for highly routine and repetitive workflow tasks. According to Jacoby, current RPA tech is capable of automating about one-third of business operations—a range that helps explain why reports from consulting firms EY and Deloitte found that RPA projects The majority either fail completely or never live up to their potential.
Instead of starting with an idealized workflow, Jacoby says AI software can be trained on what a company actually does in real-world situations. Instead of adhering to an overly standardized and rigid template, AI can understand what the right workflow is for that particular situation.
He gives the example of drafting and sending a non-disclosure agreement as part of a commercial agreement. With today’s RPA, a company can make a rule that any deal over $100,000 requires that the counterparty be sent an NDA. The software robot automatically handles the workflow of filling the NDA template for that particular deal.
The problem with such rules, Jacoby says, is that they’re too complex to capture real business logic. How about a contract price of $98,500? Most businesses will probably still want to be covered by NDAs even though this is below the threshold set for robots. Modern LLM-based software involves a lot of professional judgment about things like past data, tacit knowledge, which requires an NDA.
According to Interlom, tasks best suited for this type of automation would include procurement and risk assessment, customer onboarding, mortgage and insurance claims processing, and import and export logistics document management.
While Jacoby says humans will be needed to act as quality control checks on Interlom’s software robots, he believes that for many processes, AI robots like the one he’s building will be outsourced. Will be able to increase the amount an employee can give. 30 times the amount of time.
Interlom, which currently employs just ten people but is hiring fast, plans to target Germany’s large “Mittelstand” – medium-sized industrial companies that are a strong part of the German economy – As its initial customer base, including plans for global expansion. To America, soon.
“Every sector of the economy will eventually be rebuilt AI-first,” said Nathan Benich, founder and general partner of AirStreet. “With their experience at Boxplot and Hyperscience, the Interloom team has a uniquely deep understanding of automating complex workflows. This makes them well-placed to build process infrastructure that leverages AI for business.” It will enhance operational productivity.