Being the emerging frontrunner comes with a few challenges. You have to shout a little louder than the competition and prove your worth a little more convincingly. In the case of technology – particularly within an established industry like eDiscovery – you also need to overcome the inertia of familiarity.
When it comes to eDiscovery platforms, the familiar choice is Relativity. It’s a solid platform with strong market share and a long heritage. It was definitely a contender when we compared options for our own long-term supplier, but we quickly found that heritage doesn’t always mean superiority.
In fact, closer examination quickly revealed (no pun intended) that the most exciting developments were coming from – you guessed it – the new kid on the block: Reveal.
Where other platforms were colouring within familiar lines, Reveal was pushing boundaries – particularly with its innovative use of AI, and its world-class user experience. (It didn’t hurt that Reveal acquired Brainspace, the same AI/machine learning technology promoted by Relativity.)
Let’s take a closer look at some of the features that convinced us Reveal was the way of the future.
Customisable, trainable AI models
Traditionally, AI is purely used to accelerate review. With Reveal, however, AI capabilities have been threaded throughout the eDiscovery process.
That includes AI model-building workflows that can be applied to early data analysis as well as review, helping to categorise and prioritise data, filtering out the “noise” and amplifying the “signal” (i.e. the relevant case material) earlier, reducing volumes and hence downstream costs very effectively.
With experience, it’s possible to build these custom AI models (on top of Reveal’s out-the-box options) remarkably quickly and easily. They are simultaneously trained by Reveal’s continuous active learning (CAL) engine, whilst subject matter experts are coding documents. Then, using the resultant predictive scores, it becomes possible to prioritise documents for analysis and review, effectively surfacing the data likely to be most relevant, first.
The models themselves are also portable and can be reused for subsequent cases (with or without further training). A pre-existing AI model with an imperfect fit can provide a rapid head start into a case or in the best case, a well-suited model waiting in the wings can provide near-instant case insights and an extraordinary (and highly cost-effective) competitive edge. At the very worst, you will discover whether a model is un-suitable very quickly; the “fail fast and move on” approach.
Content enrichment, enhanced review, and reports
Reveal also uses AI to power some really effective Review Accelerators, which have proven firm favourites with our clients. They include:
- Automatic transcription of audio and video files into searchable text
- Automatic descriptive image labelling to help search and filter images
- Customisable language translations, supporting over 75 different languages during the language translation process
- A flexible interface for a customisable review experience
These are supported by some very useful review performance indicators, including reports on review accuracy, reviewer comparisons, tag sets and reviewers, and reviewer efficiency. (Reveal 11 adds reports that analyse custodian, document, and user related activities as well.)
World-class data visualisations
It’s not just Reveal’s AI innovations that impressed us. We continue to be blown away by the interactive data visualisations the platform offers, as well.
Its intuitive dashboards make it nearly effortless to filter out low value content by time, custodian or file type and zoom in on potentially relevant documents to quickly uncover hidden insights and connections.
Cluster Wheel visualisations, conceptual search and the Brain Explorer, group and interrelate large volumes of data by concept, helping to surface key topics of interest, isolate and exclude low value clusters, and explore potentially responsive documents in context.
There are also Communications visualisations which provide a map of who was discussing what with whom, including relationship and participation-level details.
Together, these tools offer extraordinary capabilities to spot and connect critical dots that may otherwise blend into the background.
The final seal on the deal for our decision to use Reveal was their obvious hunger for innovation and dedication to ongoing improvement. The platform’s developers have been hard at work releasing regular improvements and new features, as well as acquiring complementary solutions to integrate and extend the overall capabilities.
Great examples are their recent acquisitions of both Technically Creative and Ligl, enabling the rollout of data connectors to expand the existing governance and compliance workflows and providing workflow, automation and legal hold functionality within Reveal 11.
These acquisitions enable automated data collection workflows between Reveal and Microsoft 365, Teams, Slack and many more, reducing overall eDiscovery timelines, minimising risks, and cutting costs.
Choosing Reveal as our eDiscovery platform may have seemed unconventional at the time, but it’s a choice that we’ve come to appreciate even more in retrospect. That said, not even Reveal’s extraordinary technology can replace the importance of people in eDiscovery.
Check back to see our next blog that explains how the amazing team at Salient harness the technology and take it to its limits.