#eDiscovery Data Translations | Explore Tumblr Posts and Blogs

techinbiz · 5 years

Text

4 Ways Chatbots Are Changing the World

Anything worthwhile starts with a conversation. And as an application of artificial intelligence (AI), chatbots have the potential to catalyze developments in industries. A smart route to automating rendered services, chatbots create opportunities to improve customer engagement and operational efficacy in businesses. From customer relations, to IT helpdesks, human resource FAQs, marketing, and sales, chatbots are now practically everywhere. In fact, the global chatbot industry is projected to reach $5.6 million by 2023. As companies like Facebook and Telegram are already providing their won platforms are chatbots, industries are fast developing chatbot applications to add value and drive change. With its growing use cases, chatbots are starting to dominate mobile consumer markets and progressive digital industries. With that in mind, here are four ways chatbots are changing the world: Finance Chatbots Nearly 75% of all bank interactions currently take place online and via mobile devices. Wells Fargo, Bank of America, and Capital One have had in-app chatbots for years now that aim to optimize the banking experience. AI-fueled chatbots provide conversational banking to customers and automate replies that are helping to personalize financial transactions. With chatbots on the frontline, bankers and consumers now enjoy multi-channel and omnipresent customer service from their banks. Mortgage and loan chatbots are key developments in conversational banking. With bots that are constantly learning, the speed of response and transactions are improving on a daily basis. As a result, the less human involvement in transactions also reduces costs, risks, and errors. Health Chatbots Chatbots are also shaping the health industry. With natural language processing, health chatbots are being used for a range of things such as discussing symptom to automating appointment bookings. Instead of scouring sites like Mayo Clinic and WebMD, the Sensely chatbot provides patients with access to symptoms and health information at their fingertips. Meanwhile, chatbots like SafeDrugBot enables doctors to immediately view pertinent drug information and any safety considerations – providing data-driven support at the palm of their hand. With instant AI-powered assistance, health professionals will be able to cater to their patients better than ever before. Legal Chatbots One of the key areas to overcome in legal tech right now is automating eDiscovery. Companies are racing to find the quickest ways to categorize documents by relevance – and as accurately as possible. With machine learning, more legal firms are investing in "predictive coding" to maximize both precision and recall when it comes to tagging legal documents. Using natural language processing found in chatbots, translation company SDL provides a Multilingual eDiscovery Solution to access foreign language case-related content. Processing and collecting electronically stored evidence in itself already presents a legal challenge. It becomes even more challenging if the evidence has been stored across international borders, the obstacles of which are explained in Special Counsel's 'Cross-Border eDiscovery: An Introduction to Cultural and Legal Obstacles'. With this technology, contextual language will no longer be a barrier and tagging and recalling evidence will just be a chat away. Dealership Chatbots Car dealerships can be stressful and complicated places for consumers to navigate. With the deployment of automotive chatbots on platforms like Facebook, car buyers can now compare car prices through a single conversation. Chatbots in the automotive industry can even answer complex dealer-specific questions, according to Martech Today. From browsing inventory, to determining car trade-in values, checking credit scores to whether they pre-qualify for finance offers, or simply making appointments, consumers are afforded valuable information 24/7.

0 notes

duaneodavila · 6 years

Text

Why Review Is Much Harder Than You Think

Linear review is alive and well and prospering. I’ve written about companies developing dashboards to better manage review like Esquify to allow for capturing and seeing real time metrics of reviewers. I recently saw Periscope — another reviewer management platform developed by Gravity Stack, an off-shoot of Reed Smith. Esquify and Gravity Stack are not alone — virtually every review shop across the country, and many firms, have developed similar products. I’ve advocated for more metrics for years, so those developments are welcome.

But do we really need all of this linear review? Is it worth the money?

Review is essential — no question. But what you review needs careful consideration and it should almost always be a small subset of what you’ve collected. To get to that subset takes thoughtful preparation and time, together with input from the client, consideration of the case strategy and your obligations.

The truth is that review done right is really damn hard. Set aside the notion that it”s easy to throw a roomful or stadium full of contract lawyers And it’s not something that is thrown together in a day or even a couple of days. Review must be very strategic, and it starts at the beginning. No, I’m not talking about the left side of the EDRM — I’m talking about the complaint.

The pleadings (or a CID or other agency request) should always form the baseline of the review. Requests for production that come as a result of the complaint/answer/counterclaims must be relevant to the claims. And careful attention needs to be paid throughout the process of identifying what data should be reviewed to respond to those requests.

Yeah, yeah, yeah, we know, you say. But are you really paying attention? The truth is that it’s very easy to comb over the details because you have a million things to do and manage as a litigator and you need to get a production out. And the new rules in 2015 made it even harder.

There are many reasons why review is hard. Here are a few at the top of the list.

VOLUME, VOLUME, VOLUME. We have so much data. Even with a targeted, thoughtful collection, the truth is that the software we use to create, store, send and receive data is not meant to be mined effectively for ediscovery. Targeted collection is a key — no question — but even afterwards, you’ll have to cull further. PRO TIP: Have an attorney work with the custodian on the interview and the collection so you can identify search terms and culling methods that can be used for the initial review. This needs to be someone who knows and understands the complaint and effective litigation strategy all the way through trial presentation. It all ties together.

Complexity of different types of information. It’s not just about email anymore. The typical collection from a hard drive or user files includes email and attachments, plus audio files from voicemail, links to videos , instant messaging, social media and more. Forget about things like having to identify and translate documents in foreign languages — that’s another piece entirely. You need to think about what from your collection will need to be produced and at what stage, and organize the data as it goes into the database.

So many issues. Every case has multiple issues and data that pertains to each. Before you load data, think about whether you can set up an early tagging structure and tag data in bulk or organize it in sets for review. You can organize it in many different ways other than by custodian (but do that too) — by RFP #’s, but categories of data, claims, types of data, time periods, etc.

Training reviewers is hard and requires thoughtful planning. I’m not a fan of so-called first pass review. I think it’s a waste of time. Why review just for relevance when you can use technology for much of that? Review should be about getting as much value as possible out of eyes on data. Organize data by issue and have your reviewers write a report about what they are seeing every day. It’s an extra 20-30 minutes of time that will keep you review directed even more than metrics about how many docs are reviewed, how many tagged by issue, etc. But being able to do that requires having good quality lawyers — whether contract or otherwise — who have litigation experience and can help determine the value of documents to issues beyond just relevance.

Technology is not an easy button. While it would be nice, the technology we have at our disposal only augments our knowledge and is as good as how it is applied and used. Date filters can be an amazing ability to limit scope of review and production, but must be done early on in the process. We’ll talk about that later as a huge piece of cooperation. AI and PC or TAR can be of huge value, but as we’ve seen in case law, you have to be up on these technologies and prepared to discuss and defend them BEFORE you start your review. And good attorneys know that really culling down a collection requires layering technologies. Keywords, date filters, junk and privilege filters, alias detection, predictive coding, clustering, etc. are all amazing tools, and each data set requires a different approach and combination to get to what you need to create a meaningful review set.

Devising a strategy and implementing technology takes time. Iteration is the name of the game. Run something, see what you get. The strategy team should be spending hours understanding the collection.

Negotiation. Yes, I believe this is an enormous piece of review. Because if you don’t negotiate parameters, form of production, how to deal with each type of data, then you don’t know what you need to do in review to make what you promised happen. A good strong ESI protocol that is specific to the case and addresses the types of data that you anticipate needing is key. If you find something new, amend the protocol. Do this early and cooperatively and you’ll be able to set up review with confidence and avoid the Chinese Fire Drill.

Review is a necessary part of litigation — no question. But what you review, and how you review it can be done much more effectively with thought and planning. Sweat the details. Revisit the bounds of what needs to be produced and when. Have a plan and make it strategic.

There are two advantages — you’ll get to what you need to use for the case faster and not spend so much time helping the other side prepare their case through production and privilege logs, and your client’s bill will reflect valuable work product. Win / win.

Kelly Twigger gave up the golden handcuffs of her Biglaw partnership to start ESI Attorneys, an eDiscovery and information law Firm, in 2009. She is passionate about teaching lawyers and legal professionals how to think about and use ESI to win, and does so regularly for her clients. The Wisconsin State Bar named Kelly a Legal Innovator in 2014 for her development of eDiscovery Assistant— an online research and eDiscovery playbook for lawyers and legal professionals. When she’s not thinking, writing or talking about ESI, Kelly is wandering in the mountains of Colorado, or watching Kentucky basketball. You can reach her by email at [email protected] or on Twitter: @kellytwigger.

0 notes