Yaniv Makover, CEO & Co-Founder of Anyword – Interview Series
Yaniv Makover is the CEO and co-founder of Anyword, the leading generative AI copywriting solution designed for marketing performance. Yaniv graduated from Ben-Guiron University with an M.S. in Computer Science and Information Systems Engineering. Prior to working at Anyword, Yaniv conducted extensive research in the fields of Machine Learning and Natural Language Processing, working directly with some of the world’s largest publishers to optimize their advertising and content to improve their performance. With the advances of Large Language Models, it was now possible to build an AI that would be able to understand which words will make the deepest impact before you publish them. This was the beginning of Anyword. Yaniv and his family live in Tel Aviv, where he oversees operations across Anyword’s New York, Tel Aviv, and satellite offices worldwide.
What initially attracted you to computer science?
I’ve been coding from age 7, so I’ve always had a strong interest. I’ve always liked building things and it’s easy to build with software.
Could you share the genesis story behind Anyword?
We founded our first company, Keywee, in 2013. Keywee is a performance marketing platform and service that helps publishers like The New York Times and The Washington Post to acquire new audiences. By working with the world’s best writers, we began to understand the true power of accurate and data-driven copywriting. Some publishers were performing better than others, driving more conversions and engagement. From there we were inspired to train algorithms to write high-performing marketing copy and subsequently we launched Anyword to share this technology with marketers worldwide in May 2021.
How has the hype around ChatGPT impacted Anyword?
ChatGPT is very good at elevating a person’s general level of writing. But an expert marketer struggles to use ChatGPT for their purposes. It’s not on-brand and it doesn't know what copy will perform best. With Anyword, professional marketers can actually use LLMs to generate high-performing copy based on relevant marketing data.
When users are 1 click away from creating 100s of variations of copy, how does the system decide what to recommend?
It’s a combination of LLM models and Anyword’s homegrown predictive performance model, with the user’s brief helping to create the most relevant copy that has the highest potential to achieve their marketing goal (conversions, engagement, reach…) for their particular use case (Facebook, LinkedIn, Google ads vs. social, vs. website copy…)
To break it down: Our system helps users to build a powerful brief. This combined with additional instructions and talking points allows our AI to generate copy. Copy variations are then scored against Anyword’s proprietary scoring model, based on billions of marketing data points.
Our business solution goes one step further, allowing clients to build their own custom AI models built on their own best-performing campaigns.
How do you fine tune content to resonate with a specific audience or increase engagement?
Users have a lot of control here. They can set up and select specific target audiences, which include demographic and pain-point data. When selected, these factors essentially become a part of the prompt that generates copy that will be more attractive to those audiences. Anyword includes several ‘Content Improver’ templates, which specifically improve ad and social copy. The system scans existing content and provides unique options to improve the score. Users can also improve their scores directly from generated responses. Finally, in the brief process as well as on generated responses, users can offer further instruction to fine tune the results they get. Users can rely on the predictive performance score panel to understand the impact their words have on engagement.
One of the more powerful functions on Anyword is the Predictive Performance Score, how can it predict how well a specific copy will convert?
Trained on 2 billion real marketing data points, Anyword’s AI looks for relationships between the words generated and top-performing marketing assets. Based on a user’s goals, Anyword’s scoring model provides a 0-100 score that’s easy to understand at-a-glance. Users may dive deeper, however, to understand the full engagement prediction for age and gender demographics, profession, and individual pain points.
The SEO community is often concerned that if Google detects AI content that it will penalize the site in search engine rankings, how do you respond to these fears?
The driving force behind Google’s search team has been about finding valuable, relevant content for searchers. And so valuable, relevant content remains the measure, whether it was written by a person, or generated with AI. It’s still early days, so we can’t be categorical, but it appears that the widespread adoption of generative AI in such a short period of time makes it difficult to identify and downgrade results for all AI generated content. Anyword does provide a built-in plagiarism detector, which ensures that no matter how copy is generated, the final copy is totally original.
There are dozens of solutions available on the Anyword platform from Amazon product description generators, to long form blog posts. Could you briefly discuss which tools are currently the most popular, and which ones you see trending upwards?
The most popular tools on our platform actually cut across all marketing functions. The Blog Wizard, which allows for easy SEO-optimized article generation is our most popular feature. It’s followed closely by the Instagram Caption generator for organic social media, and the Facebook Ads Primary Text generator, which supports performance marketing.
The trend, especially recently, has been for massively increased usage of our Writing Tools suite. This includes the ability to improve existing copy, summarize ideas, build bullet points, expand sentences, and more. We also see more users choosing to generate their brief directly from a live URL. These are both tools that can be used together with any area in marketing, and we believe shows an increased willingness on the part of marketers to trust and work collaboratively with generative AI.
What is your vision for the future of Generative AI?
Generative AI will quickly become a part of everybody’s daily life and work processes. We’re already seeing discussions around integrations into the Google or Microsoft Office suite, Bing Search, and assistant tools like Siri and Alexa. As time moves on, we could see written works being generated specifically for individual audiences, pitching language to reading level or cultural context. Additionally, we will continue to see more and more powerful purpose-built platforms like Anyword that are built with a specific functionality in mind, and whose power will increase exponentially.
Thank you for the great interview, readers who wish to learn more should visit Anyword.
- NVIDIA: From Chipmaker to Trillion-Dollar AI Powerhouse
- Laura Petrich, PhD Student in Robotics & Machine Learning – Interview Series
- Liquid Neural Networks: Definition, Applications, & Challenges
- Patrick M. Pilarski, Ph.D. Canada CIFAR AI Chair (Amii) – Interview Series
- AI Leaders Warn of ‘Risk of Extinction’