E-commerce marketplaces make use of a number of marketing channels like emails, push notifications, etc. to reach their users and stimulate purchases. Personalized emails especially are a popular touch point for marketers to inform users of latest items in stock, especially for those who stopped visiting the marketplace. Such emails contain personalized recommendations tailored to each user's interests, enticing users to buy relevant items. A common limitation of these emails is that the primary entry point, the title of the email, tends to follow fixed templates, failing to inspire enough interest in the contents. In this work, we explore the potential of large language models (LLMs) for generating thematic titles that reflect the personalized content of the emails. We perform offline simulations and conduct online experiments on the order of millions of users, finding our techniques useful in improving the engagement between customers and our emails. We highlight key findings and learnings as we productionize the safe and automated generation of email titles for millions of users.

Recently, there's been a lot of excitement about ChatGPT – the public preview release of OpenAI's chatbot powered by the GPT3 language model. There's no better way to get people interested in – and perhaps worrying about – artificial intelligence (AI) than showing it in action. And ChatGPT certainly acts as a powerful demonstration of what AI can do today. Ask GPT to answer a question, or to create a piece of writing, and it will respond in well-structured, natural-sounding human language that most people simply would not guess has been created by a machine. Of course, this has immediately got people asking what the implications are for us humans – and one of the first professions to fear that they could be facing the chop are marketers.

As just about every aspect of computing is being transformed by machine learning and other forms of AI, companies can throw intense algorithms at existing CPUs and GPUs. Or they can embrace Graphcore's Intelligence Processing Unit, a next-generation processor designed for AI from the ground up. Capable of reducing the necessary number crunching for tasks such as algorithmic trading from hours to minutes, the Bristol, England, startup's IPUs are now shipping in Dell servers and as an on-demand Microsoft Azure cloud service. Read more about why Graphcore is one of the Most Innovative Companies of 2020. Ever tempted to click on the exciting discount offered to you in a marketing email?

A new report finds that many publishers impressed by artificial intelligence generated writing and similar technologies are doing double-takes on its price tag. Essentially, 47% of key players in publishing surveyed say they like what they see, but they're uncertain that AI can deliver a solid return on investment. The irony, according to the researchers: Other evidence gleaned from the same survey shows even minimal investment in AI brings back significant sales increases for publishers. That additional investment coin could involve little more than hiring AI-savvy employees, or training existing employees in AI. "Those who have invested in artificial intelligence are happy with their experience and will continue to invest on all levels," the researchers observe. The study, "The Future Impact of Artificial Intelligence on the Publishing Industry" -- released Gould Finch and Frankfurter Buchmesse -- is available for free download.

If you were hoping for one, definitive answer to this question, you may need to readjust your expectations. The truth is in the nuance. Every subscriber, and their relationship with your company, is unique. What's more is that every person behaves differently throughout their day. We all have different schedules, priorities, and distractions that often dictate when and where we check our email. If there was some magical time to send email, then your subscribers' (as well as your own) inbox would become flooded at that time every week. A better question to ask would be, "is there a best time for our company to send email given what we know about our subscribers?" Like all marketing channels, we must take specific steps to understand what works best for our audience. Some marketers prefer to use a test-driven approach for learning this, while others use buyer personas and qualitative feedback to optimize send time. There are even a few tools that use artificial intelligence to look at past email engagement, and customize future send time for you. Rather than try to discover individual best times, we asked marketers from the Databox Partner Program about the strategies they use to find the best time for any sales or marketing email. We grouped the responses by strategy so you can learn a range of ways to execute each one. Some of the participants even shared why they do not believe send time is a major factor, and instead focus on other parts of the email. Some tools can change the send time of each individual email based on past engagement history. Here's how this works- if you do not already have extensive email engagement data, the tool will send your emails at randomized times over the course of a few weeks to build a profile. Then, the tools will begin to personalize the send time based on past opens and clicks.

Consumers have a classic love-hate relationship with their email that financial marketers must recognize in framing their messages, research reports confirm. People often appreciate the news of offers that they receive from banks and other retail emailers. But they may cringe when checking their inboxes and see the volume of unopened messages. Reactions range from sighs to profanity to the out-loud chant "delete, delete, delete" as they render message after message forever unread. But consumers' mixed feelings about their personal emails, and their work emails as well, represent only part of what bank and credit union marketers must account for.

AI solutions such as automated emails and predictive analytics can act as a force multiplier for B2B marketing teams. Megan Wharton is an excellent sales assistant. She's been at her job only a few months, but she's easily the best person at qualifying promising leads on her team. Megan has a remarkable knack for chatting with potential customers over email, and knows exactly what to write to get prospects talking with a salesperson. While her counterparts disqualify leads after a few attempts, she keeps pushing.