Stoke Talent, the freelance recruitment platform, has tapped Amazon’s NLP suite Comprehend to create a tool to help companies handle compliance issues.
This is not the first legal tech-related tool based on Amazon’s NLP suite. For example, Search Acumen also went this route to make an application for property transaction data.
The startup told Artificial Lawyer that its Worker Classification Engine acts as a ‘screening solution’ that is able to offer early detection of ‘high-risk independent contractors and freelancer relationships’.
It provides companies – including those looking for legal staff – with a ‘continuous, automated solution that monitors and tracks each individual relationship [to a company] for complete compliance with all workforce classification laws,’ they said.
Stoke, which is based in Palo Alto, lets you consolidate matches from millions of talent profiles worldwide into one pool that you can filter by employment type, skills, location, and that can then be matched to a company’s hiring policies.
The company added that there had been a surge in freelancers during the pandemic and so their tool was especially relevant today. Getting compliance here mattered, they noted, because incorrectly engaging with a freelancer had social benefit payments and tax contribution implications for companies.
The company explained to this site that they had trained up the Amazon NLP system using ‘thousands of employee classification cases in the US’ and they had also been inputting plenty of synonyms for key terms to help the software identify cases where terminology had varied.
Interestingly, they noted that a lot of the language variation to describe freelancers comes from the multiple US Government agencies whose remits overlap with this area.
‘The federal government, state governments, and government agencies have begun to develop legislation that would ensure companies are unable to falsely classify their employees as contractors and, by doing so, prevent them from associated social benefits and tax contributions. Threats of penalties from the U.S. The Department of Labor (DOL), Internal Revenue Service (IRS), and other agencies – not to mention class-action lawsuits and failed audits – have also added a sense of urgency to the issue.
‘Moreover, each involved government agency applies its own logic to the determination of workforce classification, creating a lack of uniformity and costly room for error. Classification tests depend on very specific and individualised details that hiring managers may not be aware play a role and legal teams have no visibility into. Before Stoke Talent, a business looking to be able to guarantee proper workforce classification was required to take on these extensive legal intricacies and consistent manual monitoring undertakings on its own,’ they explained.
Shahar Erez, CEO and Co-founder of Stoke Talent, added: ‘Freelancers are key to productivity, especially in our current global economy. We are aware of many legal counsels’ concerns over workforce classification, so we’ve developed a simple, continuous, highly accurate and user-friendly solution that removes any related risk or looming concern.
‘Stoke can help companies to rely on ICs and freelancers to accelerate time to market and increase their efficiency by eliminating all concerns regarding legal and tax compliance.’
Amazon and Legal Tech (Once More…)
As noted, this is not the first legal tech-related tool based on Amazon’s NLP suite. For example, property group Search Acumen also went this route to make an application.
Meanwhile, as mentioned previously on AL, Amazon has already built a pre-trained Medical language subset of Comprehend, which it markets on its website – here.
Now, the next question has (as always) got to be: will Amazon ‘do legal tech’…? The answer is: there is no reason not to. The medical world is highly regulated and Amazon was not scared off from selling the NLP kit mentioned above. Doing the same for the law would not be a massive stretch….if they could be bothered to do it, or perhaps saw synergies with other products they sold – and they haven’t yet got into selling their own legal forms and templates, for example.
And, it’s worth mentioning that Amazon is so large now that it could launch into custom-fitted NLP sets across multiple sectors and hardly even notice it. After all, in 2020 Amazon generated total net sales of approximately $125.5 billion. Growing Amazon Comprehend would clearly not be a massive strain on the business.
That said, at the end of the day, companies such as Kira Systems have been around for over a decade, training up their NLP capabilities on commercial legal terms day-in and day-out, improving their output every year. Why buy an off-the-peg ‘vanilla’ NLP suite, or for that matter use any of the open source solutions already available if you are a law firm?
Cost is one potential reason. But, the answer is really around wanting to build and own your own tool, especially if your aim is to create a commercial product.
Although, that said, Kira is also used by at least one legal tech company that is based upon its software – see Document Crunch. So, it’s not a scenario of absolutes.
Plus, would it really matter? Lots of legal tech companies use open source NLP toolkits as a base tech layer already. The real issue is: does a toolkit like Amazon’s get so good that you could go direct to them? I.e. Comprehend become a complete end product, not just a toolkit, and you can then skip the other vendors and just use Amazon? That’s a whole other question.
So, will Amazon ever release their own trained up sub-set of legally-focused NLP software that was so good you could just use it directly, out of the box, and not need other vendors? It’s possible. And, perhaps over the very, very long-term perhaps even inevitable, given how Amazon’s strategy appears to be ‘to eat the world’.
How much actual impact it would have may be minimal though, if it ever happened. In a very specialised market full of established competitors even Amazon might find it difficult to take market share. After all, why would a law firm that had been using X legal tech specialist company for many years move to an Amazon own-brand product? Time will tell.
P.S. If you are curious about what Amazon Comprehend can do, check out the company’s info below:
‘Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to find insights and relationships in text. No machine learning experience required.
There is a treasure trove of potential sitting in your unstructured data. Customer emails, support tickets, product reviews, social media, even advertising copy represents insights into customer sentiment that can be put to work for your business. The question is how to get at it? As it turns out, Machine learning is particularly good at accurately identifying specific items of interest inside vast swathes of text (such as finding company names in analyst reports), and can learn the sentiment hidden inside language (identifying negative reviews, or positive customer interactions with customer service agents), at almost limitless scale.
Amazon Comprehend uses machine learning to help you uncover the insights and relationships in your unstructured data. The service identifies the language of the text; extracts key phrases, places, people, brands, or events; understands how positive or negative the text is; analyzes text using tokenization and parts of speech; and automatically organizes a collection of text files by topic. You can also use AutoML capabilities in Amazon Comprehend to build a custom set of entities or text classification models that are tailored uniquely to your organization’s needs.
For extracting complex medical information from unstructured text, you can use Amazon Comprehend Medical. The service can identify medical information, such as medical conditions, medications, dosages, strengths, and frequencies from a variety of sources like doctor’s notes, clinical trial reports, and patient health records. Amazon Comprehend Medical also identifies the relationship among the extracted medication and test, treatment and procedure information for easier analysis. For example, the service identifies a particular dosage, strength, and frequency related to a specific medication from unstructured clinical notes.
Amazon Comprehend is fully managed, so there are no servers to provision, and no machine learning models to build, train, or deploy. You pay only for what you use, and there are no minimum fees and no upfront commitments.’
1 Trackback / Pingback