CLM company Ironclad has formally launched a much-anticipated suite of document analysis AI tools, plus enhancements to existing capabilities – built upon Google’s ‘infrastructure’. However, and importantly, CTO and co-founder, Cai GoGwilt, told Artificial Lawyer ‘the Ironclad team has developed and trained our own models and functionality’ for the NLP tools.
This point about who is making the finished NLP product for doc analysis is central to any discussion of what part Google may potentially play in disrupting the legal tech world.
As Google told this site back in November 2021 – see interview – it has developed a ‘Contract DocAI’ capability of its own via its Google Cloud group, which for now is only working with a very limited number of other companies to see how it plays out.
If Ironclad is however ‘finishing off’ the training and adapting the ‘vanilla-ish’ Contract DocAI NLP provided by Google, then a number of legal tech companies can perhaps wipe the sweat from their anxious brows. Because if Google’s DocAI isn’t going to be used ‘directly’ for real-world client use on legal docs without any training from a third party, then legal tech companies don’t have too much to fear – yet. In fact, Google’s NLP ‘kit’ then could be viewed as the same as the other vanilla text analysis offerings, (some open-source, others paid for), that already exist in the market.
What people in this sector were potentially freaking out about was the prospect that Google’s DocAI would be ‘client grade’ at the point of sale, ready for instant application to doc analysis on whatever customers wanted it to do, and available to all at a reasonable price.
Twin that with all of Google’s many other offerings, and you’d have a really compelling product that would make most NLP-based doc analysis tools not exactly redundant, but they would be facing incredibly tough competition. However, for now, that doesn’t seem to be the case. This is more of a ‘Google DocAI Inside’ type approach, (after the famous advertising slogan by Intel), as well as Ironclad leaning on the wider Google Cloud infrastructure – but, not Google taking the lead itself on the final product.
What’s On Offer?
So, with that said, what has Ironclad rolled out? Plus, what is totally new and what is an advance on what was there already? Here is what they say. First, the ‘enhancements’:
‘Today, Ironclad has announced enhancements to Ironclad Smart Import, an AI-powered contract migration tool that uses optical character recognition (OCR) to scan, index, tag, and store contract data at scale. With Smart Import, customers can:
- Make contracts full-text searchable: Industry-leading OCR technology makes all of your contracts full-text searchable, including scans of wet-signed contracts.
- Confidently automate data extraction: Ironclad AI extracts key terms and clauses with over 90% accuracy.
- Use your contracts and data instantly: Make your contracts and data fully searchable and reportable, instantly.
- Control contracts with AI-powered review and negotiation
Meanwhile, they have also launched in beta, Ironclad Playbooks, ‘which uses AI-powered clause detection that alslows customers to review and negotiate contracts more intelligently’.
With Playbooks, companies can:
- Automatically analyze and intelligently negotiate contracts, faster: Ironclad AI automatically analyzes contracts, flags areas that require a thorough review, and provides suggestions on how to negotiate based on legal-approved guidelines.
- Review third-party paper at scale: Ironclad AI automatically loops in the right approvers at the right times to streamline negotiation, enhance contract collaboration and compliance, and speed up contract review times.
- Convert PDF files to DOCX for contract reviews: Ironclad now offers the ability to instantly convert PDF files to DOCX when editing documents to slash the time and effort required to get contracts uploaded, reviewed, and approved.
So, there you go. A mix of improved tools and new tools, ‘based on Google infrastructure’, but with Ironclad doing a lot of its own training to produce the final, polished product that will be used by the clients.
The Interview
Artificial Lawyer caught up with CTO, GoGwilt, to find out some more. Here is the interview below:
Why do this now? Why not earlier?
There are two main reasons we’re releasing Ironclad AI now: (1) NLP technology has only recently become viable in the contracting space, and (2) Ironclad’s platform has expanded to set a foundation for Ironclad AI.
As soon as NLP capabilities became acceptable for legal contracts (with the advent of Transformer models), Ironclad started work on Smart Import, our industry-leading contract data extraction capability. When we released Smart Import, not only were we building out this best-in-class contract data extraction tool, we were also investing in the underlying foundation of Ironclad AI.
With that Ironclad AI foundation established, and battle-tested with Smart Import, we were able to quickly move forward with bringing contextual intelligence to the entire lifecycle of contracts.
How much did Google play a role in these capabilities that you are offering? And how much of the new products are in effect ‘With Google DocAI inside’?
Google AI and Google Cloud helped play a foundational role for our AI, acting as the infrastructure which Ironclad AI runs on – similar to the way many companies run their software on top of AWS, for example.
Using this as our infrastructure, the Ironclad team has developed and trained our own models and functionality. So in essence, these new products are built by Ironclad, powered by Ironclad AI, and supported by Google Cloud.
How do you think this will compare to other NLP / doc analysis offerings across the CLM field? (As there are plenty out there now.)
Everything we do as a company is to help our customers accomplish three things: speed up their businesses, drive results, and operate more effectively. This is why we’re integrating AI across our entire platform and the entire contract lifecycle – from creation to analysis – rather than simply in one area like document tagging. Our approach is different from most in a few areas: first, we provide AI-driven insights directly in context for the decision-maker.
Playbook suggestions are made within the document itself, and Smart Import shows exactly where in the contract a piece of data was extracted from. Second, we make it easy to get started with powerful functionality. Users can control changes and suggestions during negotiation, select how to use predicted values in the Repository, and everything works well out-of-the-box; no need for a lengthy on-boarding, training process, etc – it just works.
Lastly, we make playbooks available for all users with editing permissions, with full legal control of course, to let companies move faster and reduce workload for the legal team – whereas most offerings focus solely on the legal team. Since contracts are a collaborative effort, we want to give legal teams the tools they need to empower the rest of their team and help their contracts move faster, more effectively.
How much of an investment has this been for the company – in financial and human effort terms?
I personally have been leading a team on this effort, with members of our product and engineering teams actively dedicated to developing and shipping these new products. Our entire senior leadership team is involved in guiding our general approach to AI, ensuring we’re being strategic, aligning across the business, and driving the highest impact and best results for our customers.
How do you see IC growing now? What are the ideal clients and how global do you want to be?
Ironclad is growing at a fantastic clip – our team is now over 500 employees-strong, we recently crossed the 1 billion contracts processed threshold and are on pace to power another billion this year alone, and we’ve brought on several key leadership hires within the last year as well. In terms of our customers, we will always have a strong focus on the legal user – but we’re seeing massive growth in other areas for our customers.
Most teams have about eight business users for every legal user, and we’ve seen spikes in usage for sales (a 285% increase in sales workflows launched YoY), human resources (230% increase YoY), and procurement (214% YoY). We plan to continue to serve our legal users as we expand into use cases where we’re seeing our customers gain massive value. In terms of our regional approach, we’ve recently added several positions working out of our EMEA remote hub and plan to continue expanding our reach outside of the U.S.
Thanks Cai and congrats on the launch!