Thoughts, observations and musings on the law, eDisclosure and various other things
Introduction – costs, confusion
At i-Lit Support we aim to be Amenable, Available, Affable, Able and Affordable.
This blog discusses the difficulties in assessing whether or not E-Disclosure service providers are ‘affordable’ and what we have done to tackle some stubborn issues that continue to exist within the industry.
After nearly 20 years of comparing the pricing models of E-Disclosure service providers on behalf of law firms it’s fair to say that we have seen every possible permutation of how service providers charge for their services. No two service providers are the same but, broadly speaking, pricing gets split into the following bands:
Not all matters incur costs at every stage and not all service providers charge for all of the services or use the same model of charging at every step. This can make price comparison (and more importantly, budgeting) almost impossible.
As we begin to add Generative AI pricing into this mix the complexities of comparing e-disclosure pricing becomes truly perplexing.
In a 2017 blog post from L21 (https://www.law21.ca/2017/06/rise-market-pricing/) Jordan Furlong states the following,
“buyers’ pricing preferences are going to include:
Roughly, I suspect, in that order.”
This struck home with us.
Law firms should be able to tell their clients how much an externally purchased service is going to cost, the pricing should be simple to understand (particularly to those outside of the industry), convenient to calculate, be consistent across matters / clients, as well as proportionate and affordable to the matter at hand.
Clearly if the original project specifications change then costs will change but if they don’t then the original price shouldn’t vary.
Currently, law firms find it hard to be certain of what their monthly costs from service providers are likely to be. Of the many hidden costs that contribute to the lack of certainty ‘time’ is the most obvious variable that service providers charge for. Rates for project management and technical time are generally around £200 per hour; it doesn’t take many such hours to radically change the initial cost estimate, which will typically not have included these hourly costs other than as an “incurred on an as-used basis” line item.
Uncertainty is neither the friend of the law firm, their clients or the service provider. Uncertainty creates mistrust, frustration and discontent for everyone involved. Constant discussions about invoices both between law firms and clients and law firms and service providers leave everyone dissatisfied.
What have we done about this?
We don’t want to be in the position of wasting our and your time having conversations with clients trying to justify pricing for,
so we don’t charge for those services.
Our pricing structure is simplified down to 2 costs.
We believe that this methodology provides the most transparent pricing available on the market today. It’s fair, it’s transparent and it is predictable. More importantly it enables us to be the most cost-effective provider in the market.
Hang on, what about Generative AI pricing?
We’ve transferred this methodology to the intricacies of Generative AI pricing. We can tell you precisely what we will charge you for each ‘question’ you ask of the document collection that you are analysing. Ask us what we will charge you, and we will tell you (and we are already offering that service today). Ask any of our competitors what they offer in the AI space today and more importantly what they charge for it and see what answer you get.
Hang on, review costs are a major element of disclosure,what about those?
We offer low-cost high-quality document review staff to any law firm; we’ll manage the process (again with no extra project management fees) and if your firm is a subscriber to our services we will offer a further discounted or fixed price review service (whichever makes better sense in your circumstances).
If you have ever had concern, disappointment, frustration or anger at E-Disclosure costs then please feel free to reach out for a discussion. We are very happy to discuss our philosophy, which is to offer a more ‘personal and pragmatic approach to E-Disclosure’.
Introduction – moving from theory to reality
In our last blog we discussed the potential for GenAI eDisclosure tools to transform litigation and investigation matters by augmenting the document review and analysis processes.
The more we use Discovery Partner the more excited we get about its potential application in the matters that we will be undertaking in the weeks and months to come; it is fair to say that that potential is breathtaking.
New technology, new approaches
The first thing you notice is the sheer speed of ‘old fashioned’ keyword searching; a demo database of over two point seven million documents produces search hits, ranked by relevance in around two seconds; the indexing alone is impressive and blistering fast compared to some other tools.
That’s impressive, but it’s just the start of Discovery Partner’s capabilities; send documents to its ‘Analyse’ folders and you can then unlock the power of multi-LLM engines to quickly gain insights and analyses that were simply not available in previous platforms.
‘Multi LLM’ may sound a little complicated, but in fact it’s a powerful feature in Discovery Partner; you can decide which LLM you to use for Summarising and Synthesising documents (i.e. analysing and reporting on them).
One other benefit of using multi-LLMs is to minimise costs; not all documents need to be summarised by GPT when Claude (by Anthropic, an AI company backed with billions of dollars fro Amazon) can do this just as well (if not better) and at a much lower cost (we’ll come back to costs shortly).
Workflows, now and in the future
Ok, sounds great – but how do you actually use it? After all, while the CPR mentions the use of ‘… technology assisted review software and techniques’, it never anticipated that technology would advance quickly to a position where AI could potentially replace human document review!
It’s unlikely a judge in England and Wales will permit a first pass document review process to be replaced by an AI-only review (although this is probably inevitable, one day); so how can I leverage the power of Discovery Partner today?
In litigation matters we see Discovery Partner being used as an amazingly powerful, insightful, early case assessment tool – identify key documents, concerns and themes in minutes, providing a list of individuals from whom witness statements may need to be obtained and providing questions to put to the client. All of this undertaken in minutes and at a minimal cost.
It would a crying shame not to use Discovery Partner for a few days or weeks at the outset of a litigation matter; the shift from ‘search and review’ to ‘questioning’ documents is going to revolutionise the practice of law.
We have been working on how best to take the knowledge gained from Discovery Partner and make this available for further review (either in Discovery Partner itself or in a complimentary technology such a Relativity) and can see this being a great way for legal teams to become familiar with the power of GenAI.
The typical response to seeing a technology as powerful as Discovery Partner is to ask ‘how do I know its finding everything’; that in a way is the wrong question, its more that GenAI is providing you with insights and avenues to explore that you may never have identified until you were a long way into your document review process. We all remember similar concerns being expressed about Technology Assisted Review, which is now fully endorsed by the judiciary.
Moving to Investigations, the sheer speed and power of analysis that Discovery Partner can apply to documents makes previous investigation approaches seem antiquated; the days of ‘search and review .. and review … and review’ are surely coming to an end. Questioning a document set rather than try to ‘get lucky’ with searches will dramatically expedite investigations and reduce costs accordingly.
This golden future of immense efficiency gains sounds almost too good to be true, yes?
Ah, but what about the costs? AI is surely expensive, after all the amount of computing power required is mindboggling!
This is where a multi-LLM approach can help to produce highly cost-effective results; GPT is an excellent technology but, as we mentioned above, there are far cheaper alternatives that can do elements of the AI process just as well. However, it is clear that any GenAI costs will be hugely outweighed by time (and associated cost) savings. If the costs for ‘reviewing’ 100,000 documents using GenAI were say, £5,000 then this would compare rather favourably with a typical first pass review cost of say £85,000 and would take a fraction of the time to complete. A cost saving of over 90% is certainly compelling!
We are in the early days of applying GenAI to matters, costs will become easier to calculate and will likely reduce as the technologies become more
Want to give GenAI a try?
We are happy to work with law firms and their clients to provide an opportunity for them to ‘test drive’ GenAI; just drop us an email and we can take it from there …
Introduction – a very brief history lesson
Having been in the disclosure management game for a combined fifty years, it takes a lot to impress Mike and I when it comes to technology.
We have seen review platforms develop from the likes of Concordance and DB/TextWorks all the way through to Relativity and, truth be told, there hasn’t been any radical change to the approach of managing and reviewing documents over that time.
Sure, the interfaces have become slicker and the ability to use threading, near-duplicates, and machine learning (to help prioritise and expedite reviews) has certainly helped the technology to keep pace with the ever-increasing volumes of data – but where is the ‘next big thing’ that is going to change the model of disclosure and the profession itself?
Is AI ‘the next big thing’?
Well, we don’t have to wait any longer .. the advent of Generative AI technologies (GenAI) is going to allow for an unprecedented level of unsupervised document analysis and review,; massively expediting and automating the review process and turning months of work into hours (this is not hypothetical, we have seen it with our own eyes and its truly stunning!)
How powerful is this technology?
How would you feel if you could simply ask questions of your documents, using simple English language questions, which then provides you with a level of analysis and insight in minutes that would previously only have been available from an exhaustive manual review of those documents?
While a manual review of say, 200,000 documents would take a team of 10 reviewers at least 50 days to complete (at a cost of around £175,000), the same analysis can be undertaken using GenAI technology in a matter of hours at a tiny fraction of the costs.
Ok – so what sort of questions could I ask of my documents?
Let assume you are involved in an investigation (although the approach is just as applicable to litigation) you can ask complex questions of those documents that not only identify relevance and key documents, but which then go on to provide you with a more fuller understanding of the potential consequences of the matter / allegation(s).
In our example of an investigation, we could ask the following question of our dataset (irrespective of the size of that document population):
“I want you to act as though you are a senior forensics and investigations professional with thirty years of experience.
Prepare an investigation report regarding the information you have found and present it using the following structure:
Section 1 : A bullet point listing of the people involved including their positions and their roles in this discussion.
Section 2 : A bullet point timeline of all the key emails you have reviewed by date ascending. Give me a two line summary of each including From and To.
Section 3: Summarise the issues involved and the pros and cons discussed for each.
Section 4 : List any rules, regulations or statutes being discussed with a summary of how each relates to the discussion.
Section 5: Offer your conclusions and recommendations”
What are the ramifications of this technology?
GenAI technology will have a profound impact on many aspects of litigation and investigation matters; a few of our initial thoughts are:
Is AI coming for our jobs? Is it a friend or foe?
While AI will have a massive impact on many areas of business, including the law, these are likely to take some time to manifest themselves; after all, some of the more radical changes to the practice of litigation matters will require judicial approval and it is likely that it will take some time for lawyers to ‘trust’ the analyses being provided by the AI.
For now, we feel that AI should be seen more as a friend (a super intelligent Associate with every single document memorised, if you like) than a foe …
We have experience of managing matters ranging from small commercial disputes up to global, multi-billion dollar litigations. We aim to provide an agile, conflict-free and personal service; leveraging state of the art technologies while delivering predictable, highly-competitive pricing.
In addition to more traditional technologies (such as Relativity and Brainspace) we also host Discovery Partner – a document review and analysis tool driven by GenAI and capable of delivering the insights and results mentioned in the article above.
If you would like to learn more about GenAI or i-Lit then please feel free to contact us.