Matthew Grant, Senior Director, Advanced Solutions, International at Epiq, joins George Socha, Senior Vice President of Brand Awareness at Reveal, for ACEDS #eDiscoveryLeadersLive.
Matt is Epiq’s senior director of advanced solutions for the EMEA and APAC regions. He leads a team of consultants who specialize in the application of advanced analytical technologies to complex datasets, driving the Early Case Assessment process. Matt and his team guide clients through best-practice strategies for the use of technology during legal reviews, deploying proven strategies for data filtering backed by defensible statistical modelling. Previously, Matt worked at Epiq as a senior director of document review services for Europe, overseeing the day-to-day operation of Epiq’s European document review centers. He has deep experience in the deployment of analytics and technology-assisted review tools during managed document reviews, and specializes in managing complex matters with challenging requirements, both on-site and across multiple jurisdictions. Matt is an Australia qualified lawyer (NSW), a European Certified Information Privacy Professional (CIPP/E, CIPM), and a member of the International Association of Privacy Professionals (IAPP).
Matt focused his discussion on the use of predictive coding and other analytical capabilities.
Key Highlights
-
- [1:04] Introducing Matt.
- [1:31] How Matt got into eDiscovery.
- [2:29] How Matt ended up at Epiq and his path to his current role.
- [3:30] The advantages of being in Perth.
- [4:07] Predictive coding in Australia: used by default by many firms.
- [5:28] Predictive coding in the UK: widely used by firms.
- [5:43] Why the differences in use in Australia and the UK.
- [6:08] Use in Australia by size of firm.
- [7:06] Defining sectors of firms in Australia: top tier, middle sector, international with a local presence.
- [9:10] Prevalence of predictive coding use by sector and by type of matter.
- [11:11] For exploratory investigations: less predictive coding, more old school with keywords and linear review.
- [11:59] Challenges with relying heavily on keywords … especially without sampling the nonresponsive documents.
- [13:53] Keywords as a beginning but not an end.
- [14:19] Predictive coding and analytics prioritization, culling, and QC workflows used at Epiq.
- [15:38] What gets his team excited: taking a discursive approach using advanced analytics tools.
- [17:00] Making review more efficient by using the best of people and technology.
- [18:37] Building a better partnership between law firm and service provider.
- [20:08] Building predictive coding models: the respective roles of law firms and service providers.
- [20:41] The role of end clients and how it varies across markets: Asia, Australia, and the UK.
- [21:37] When regulators get involved.
- [23:35] Types of tools or techniques avoided or net yet used.
- [25:26] Two generative AI use cases: kick-starting review and assisting with legal research.
Key Quotes
- “In the Australian market, I have found that predictive coding in particular, but the use of analytics and TAR more generally, is quite far advanced. Most firms will use predictive coding for document review by default…”
- “In the UK, [predictive coding] is widely used, but I would say there are still plenty of firms which are still dipping their toes in the water and it’s probably not as embedded as I have found in Australia.”
- “The inhouse [predictive coding] capabilities belong to probably the top five to ten [Australian firms] … and they are well used and there’s some fantastic expertise in those firms…. The sense I get speaking to them is that they are heavily used for litigation in particular, certain kinds of regulatory inquiries, but perhaps not for investigative inquiries.”
- “One of the nice things about predictive coding technologies and technology assisted review technologies more broadly is that you’re not reliant on your best guess at what keywords need to be.”
- “My preference, the thing we recommend to our customers, is that if you’re going to use keywords, use them to find a small set of documents that you know are relevant, that you know are responsive, and then let’s put those into the model because it’s likely to do a better job.”
- “The stuff that really gets my team excited is the more discursive stuff. It’s where a client comes to you and they say, ‘We have a population of documents and these are the two, three, five things that we’re trying to figure out – can you help us with that?’”
Connect with Matt