With the release of the new text variants suggestions feature, Berlin-based AI company Retresco, an expert in Natural Language Generation, provides a software function that uses artificial intelligence to automatically produce phrasing suggestions in the form of complete sentences. This function automates the creative writing process, shortens it, and increases text variance without user intervention, augmenting human creativity with AI.
Retresco’s NLG software (textengine.io) independently suggests high-quality texts based on a sample sentence entered within milliseconds and without complicated setup by the user. The automatically generated texts can be adopted either partially or completely and be revised at any time. Human and machine thus work hand in hand: the software generates data-based text suggestions, while the human user ultimately decides how to use them. The result: significantly greater text variance, more efficient processes in the creation of texts, a better user experience, and support for the most difficult aspect of writing: creativity. This feature is particularly relevant in cases where large volumes of versatile, high-quality text are required at frequent intervals and often under time pressure. In e-commerce, for instance, online stores need numerous product descriptions that, above all, have to be highly varied and SEO-optimized.
To help brands build better digital experiences and establish greater digital trust with customers, Contentsquare, a provider of digital experience analytics, announced a an analytics solution that allows teams to access critical revenue insights without any use of cookies. This announcement comes on the heels of a recent $500M Series E round led by Softbank. Contentsquare has never relied on third-party cookies, and instead aggregates and analyzes trillions of consumer interactions that demonstrate intent, such as mouse movements, touch and mobile interactions, to help brands deliver the best possible experiences to their customers. The solution will now give businesses the option to turn off both first and third-party cookies. This latest innovation extends Contentsquare’s privacy-first approach.
Language I/O released its multilingual chatbot for Salesforce. Customers expect immediate responses from businesses when they are searching for help with an issue, and 64% of customers expect 24-hour customer service support to be available. The challenge comes in meeting those customers in their native language, so effective problem solving and direction is provided. In the US alone, approximately 13% (50 million people) of consumers are not native English speakers. Messages passed to chatbots, as with live chat, are often not written out in complete sentences with proper grammar. Customers are more conversational, the way people actually speak to each other in real life. The challenge of multilingual translation for chatbots, that Language I/O solves, is processing all the messy user generated content (UCG) including jargon, misspellings, acronyms, product names, etc.
This week I suggest articles by Walid Saba and Joshua Benton and have news from Google, TeamViewer and SAP, Amplitude, Jorsek, Asana, and SpeechLive. (<2 min)
You’ll note that I continue to experiment a bit with the format. After I try out a few more things I’ll follow up with the long-promised survey, especially given Apple’s new Mail Privacy Protection. (See Joshua Benton’s article below.) In the meantime just reply to this email to let me know what you think.
Opinion / Analysis
Walid Saba
Ontology, knowledge graphs and NLU: three pillars of one and the same system
The application of enterprise knowledge graphs and natural language understanding and processing continue to grow, for good reason, but neither is easy and the combination even less so. In this short piece Walid Saba identifies a key problem yet argues that this combination, plus ontology is, in general, necessary for success. How to accomplish this? Well, he’s not the only one looking at this. The company he works for, Ontologik, is in stealth mode, but their site has links to an accessible presentation, and to more technical research.
Joshua Benton provides a good summary of last week’s announcements publishers big and small will care about. It’s a mixed bag, but changes to notification controls and the coming end of open rate statistics for newsletter publishers, like us, will not be pleased. The only thing we track is activity in our ad-free newsletter which provides important customer feedback.
The Gilbane Advisor is curated by Frank Gilbane for content technology, computing, and digital experience professionals. The focus is on strategic technologies. We publish recommended articles and content technology news weekly. We do not sell or share personal data.
DataStax announced Astra Streaming, a scalable, multi-cloud messaging and event streaming platform built on Apache Pulsar. Astra Streaming is integrated with DataStax’s marquee serverless database, Astra DB, to deliver a multi-cloud solution for managing both data in motion and data at rest. With the introduction of Astra Streaming, DataStax aims to deliver on its vision of an open data stack for today’s multi-cloud applications that require massive scale, zero-downtime availability, and high performance. Astra Streaming Features:
Global scale, cloud-native streaming, powered by Apache Pulsar without the complexity of self-managed solutions
Compatible with Apache Kafka and Java Messaging Service
Multi-cloud
Simple developer APIs for streaming
Handles high-volume queuing and pub-sub messaging and more complex messaging patterns
Pay-as-you-go pricing
Astra Streaming is available today in a beta version. To get started with Astra Streaming, create a free account.
In its second major release of the year, DataRobot announced several product upgrades to its Augmented Intelligence platform designed to further democratize AI. The 7.1 release introduces:
MLOps Management Agents – DataRobot’s MLOps Management Agents provide advanced lifecycle management for an organization’s remote models. Management Agents understand the state of any remote model regardless of how they were created or where they are running, and can automate various tasks.
Feature Discovery Push-Down Integration for Snowflake – Joint DataRobot and Snowflake customers can benefit from the automatic discovery and computation of new features for their models directly in the Snowflake Data Cloud.
Time Series Eureqa Model Enhancements – DataRobot Automated Time Series now runs its unique Eureqa forecasting models as part of the regular Autopilot process. Eureqa models are based on the idea that a genetic algorithm can fit different analytic expressions to trained data and return a mathematical formula as a machine learning model.
No-Code AI App Builder– the No-Code AI App Builder allows customers to quickly turn any deployed model into a rich AI application without a single line of code.
Additional product upgrades: Data Prep for Time Series, Nowcasting for Time-Aware Models, Automated AI Reports, and Prediction Jobs and Scheduling UI.
LivePerson, Inc., a provider of conversational AI, and Adobe announced an integration to help brands transform digital customer experience by seamlessly extending personalization from digital experiences into messaging channels and one-to-one conversations at massive scale.
The LivePerson-Adobe combination focuses on the seamless integration of LivePerson’s Conversational Cloud with Adobe Experience Cloud to help brands serve highly personalized and contextualized messages, recommendations, and offers on their customers’ favorite messaging channels. The integration helps brands to:
Enrich customer intelligence, track attribution, and drive long-term customer value by easily sharing intent data from customer conversations on the Conversational Cloud with insights on that customer’s activity and history from Adobe Analytics
Capture more qualified leads and accelerate net-new conversions by integrating the Conversational Cloud with Adobe Marketo Engage for proactive, automated conversations
Increase sale conversions and decrease checkout abandonment by offering relevant assistance at any point in the buying journey
Apply LivePerson’s Natural Language Understanding, built on over 20 years of goal-based customer conversation data, to these engagements
The Decentralized Identifier Working Group has just published a second Candidate Recommendation Snapshot for the Decentralized Identifiers (DIDs) v1.0.
This document defines Decentralized identifiers (DIDs), a new type of identifier that enables verifiable, decentralized digital identity. A DID identifies any subject (e.g., a person, organization, thing, data model, abstract entity, etc.) that the controller of the DID decides that it identifies. In contrast to typical, federated identifiers, DIDs have been designed so that they may be decoupled from centralized registries, identity providers, and certificate authorities. DIDs are URIs that associate a DID subject with a DID document allowing trustable interactions associated with that subject. Each DID document can express cryptographic material, verification methods, or services, which provide a set of mechanisms enabling a DID controller to prove control of the DID.
Candidate Recommendation means that the Working Group considers the technical design to be complete, and is seeking implementation feedback on the document. The group is keen to get comments and implementation experiences on this specification as issues raised in the document’s Github repository. The group expects to satisfy the implementation goals (i.e., at least two, independent implementations for each of the test cases) by July 17, 2021.
