One of the greatest strengths of Reveal Ask as a GenAI tool for eDiscovery is that rather than being a stand-alone tool, it is fully integrated with the rest of the platform. Both upstream and downstream.
Upstream integration of Reveal Ask as a GenAI tool for eDiscovery
We’ve discussed basic filtering using date ranges and keywords, but of course all the other capabilities of Reveal are available to help you make your queries and the context you send to Ask more granular, either in isolation or combining several techniques.
Those would include:
- Communications analysis. For example, you may have filtered around key terms but then use the communications view to further limit the context to the relevant communicators or perhaps just the domains of interest. Or perhaps by viewing the communications chart you may expose further persons of interest that you were not aware of, that warrant deeper scrutiny or inclusion in the investigation.
- Conceptual search. Rather than using pure keywords, you may want to explore other related concepts or terms that occur regularly in proximity to the initial keywords or terms. And using the Brain explorer to do so graphically. If you limit the exercise purely to keywords, you may miss critical, relevant content that uses synonyms and certainly restrict yourself from exploring those parallel concepts that may also be worthy of consideration.
- Cluster Wheel. The cluster wheel, when combined with Ask, provides an excellent way to cast the net a little wider before culling back to the more relevant content, to improve the likelihood that you have considered the most relevant content in any investigative process. By considering lexically similar terms and commonly related vocabulary to the initial search terms (keywords), you may wish to submit the entire cluster to Reveal Ask to check if the resultant references include relevant content from the estate that would otherwise have been missed.
- Heatmaps. Use of the matrix-based charting of ‘hits’ by your two chosen dimensions, may be a helpful strategy for refining the set of data to which you expose Ask.
Downstream integration of a GenAI tool for eDiscovery
Having run some queries in Ask, the logical question is “what next”?
- Conventional Review. The obvious strategy would be to pass the results into the conventional grid view for review, or foldering them for later review once you have built up content for all the themes or angles you wish to explore in the matter. This is probably where most people will end up in the first instance.
- Seeding Machine Learned (ML) models. Having used Ask to surface likely relevant content (and remembering the discussion above that you cannot be sure that Ask has found or used all relevant content), one approach we have explored is using the results to seed a machine learned model. By reviewing the seed set for relevance, the resulting model can then be applied to a wider corpus to check for other likely relevant content. As with any ML model, the speed at which it stabilises will be a function of the richness of the dataset, so this approach may take several rounds of training, but particularly on larger data sets, it is a strategy we believe may be helpful. Or one that can be applied as a QC step to try to establish that all relevant material has been discovered, a little like elusion testing.
The essential point to remember is that because Ask is fully integrated into the Reveal platform, the process can be bi-directional and iterative, with Ask informing and being informed by the rest of the platform. How you build this capability into your strategy will likely vary, matter by matter, based on the specifics of the case and clearly there is a skill in how you get the best from the tools, which will come with experience.
GenAI for eDiscovery: Success stories using Reveal Ask
We’ve been testing out RevealAsk with our clients to evaluate the potential for GenAI in their eDiscovery projects. Find out our successes, what we’ve learnt along the way and our views on where the technology will take us next.
1. Introduction, a reminder about how GenAI works in eDiscovery and a practical example about using Reveal Ask.
2. The art of prompt engineering and what we’ve learnt about the importance of getting the prompt ‘right’.
3. How to use Ask to accelerate an investigation: our learnings and strategies for success.
4. Getting the most from the integrated tool set – the power of Ask in combination with the wider Reveal toolset.
5. What works well, which use cases are best suited to the Ask capability and some observations about future enhancements
