Ivo founder Min-Kyu Jung: "It's still possible to build great companies from New Zealand but, empirically, it seems to be much harder."
Ivo founder Min-Kyu Jung: "It's still possible to build great companies from New Zealand but, empirically, it seems to be much harder."
A startup founder gives candid replies to questions on whether AI can trusted not to make mistakes, and whether the technology is a threat to lawyers’ jobs.
Former Bell Gully solicitor Min-Kyu Jung (29) relocated from Auckland to San Francisco, where he’s just raised US$4.8m in venture capital for hisseven-person, 1-year-old law-tech start-up Ivo.
He describes Ivo as a “generative AI-powered contract review solution”. Like the Tauranga-based, US-focused LawVu, its target market is big organisations’ in-house legal teams, rather than law firms.
“Contracts are foundational to commerce, but reviewing them is a key bottleneck for important business processes,” he says.
“By using AI to reduce the time, effort, and cost of negotiating contracts, we’re making it easier for businesses to work together.”
Ivo's team, with Min-Kyu Jung at left. The plan is to triple the numbers to 20 over the next few months.
Early customers include large US outfits like Carparts.com, SCAN Health, WordPress owner Automattic and e-Health firm Commure.
The new funding will go toward automating more legal processes with AI.
The seed round was led by two Californian VCs, Uncork Capital and Fika Ventures, and supported by three previous investors: Kiwi VCs GD1 and Phase One, plus Daniel Gross - a former Y-Combinator partner who worked on AI projects for Apple for five years after the tech giant bought his search engine, Cue, for US$40m-plus in 2013.
The start-up only began commercial sales a few months ago and has yet to reveal any sales figures. It began life in New Zealand as “Latch” but after finding a NYSE-listed firm of the same name, switched to Ivo after Ivo of Kermartin, the patron saint of lawyers.
The Auckland University grad answered questions from his new home in the Bay Area.
Herald: What’s your message to legal teams worried about AI “hallucinating” or otherwise making errors?
Jung: Hallucinations happen because a large language model [a generative AI like ChatGPT, Facebook’s Llama and Google’s Bard - now Gemini] will confidently answer a question even if it can’t find the answer in its training data.
Hallucinations can be mitigated in two ways: first, by giving the LLM [large language model] access to additional data and giving them tools to request more information if necessary; and second, breaking the problems they’re solving down into their component parts so the LLM can focus on one issue at a time.
Ivo uses generative AI to review contracts. Jung says the average lawyer spends between 40 to 60 per cent of their time drafting and reviewing contracts. He says Ivo can shave 45 minutes from the time required to review the average contract for a 75 per cent efficiency gain.
Ultimately, large language models [LLM] are probabilistic and will still sometimes make errors.
Lawyers should be expected to independently verify LLM outputs in the same way they might verify the work product of a junior lawyer before giving it to a client.
That doesn’t mean there still isn’t enormous utility in having a junior lawyer do an initial review of a contract.
What’s your message for lawyers or law clerks etc worried about AI taking their jobs?
LLMs, like all technologies, provide leverage - lawyers who are proficient at using LLMs will find they’re able to produce higher-quality work more efficiently than they would otherwise.
Overall, I think this is a positive development. Most lawyers I speak to aren’t particularly fond of spending their evenings doing low-level contract work like preparing issues lists, making redline adjustments, or verifying that defined terms are being used correctly. LLMs will free lawyers from mundane tasks and allow them to focus on being a trusted adviser to their clients.
Was there a specific moment, while working at an established law firm, that gave you inspiration for your start-up?
No specific moment, other than a general dissatisfaction with the status quo and feeling a strong compulsion to do something about it.
Why move to the US?
The San Francisco Bay Area is the best place in the world for an AI start-up to be based right now.
We moved our offices from Auckland to San Francisco recently, including our staff. All new staff will also be based in our San Francisco office.
It was important to us that we were surrounded by the best AI talent in the world.
We think we’re at the very frontier of pushing LLM capabilities; there aren’t actually that many people in the world who can help us with the problems we’re solving, and a wildly disproportionate percentage of them are based in the San Francisco Bay Area.
We also think there are other important ancillary benefits of being based here – the founders in the Bay Area are encouraged to be more ambitious and it’s easier to raise capital.
It’s still possible to build great companies from New Zealand but, empirically, it seems to be much harder.
Chris Keall is an Auckland-based member of the Herald’s business team. He joined the Herald in 2018 and is the technology editor and a senior business writer.
