In the age of generative AI, corporations have generally used prompts at every step of a content generation process. AI startup Writing thinks there’s a greater way.
The San Francisco-based startup was founded in 2023 with the goal of enabling highly customized, branded generative AI content. Typeface has built integrations for its technology with multiple vendors, including Microsoft and Google, to assist corporations with marketing workflows.
Today, the corporate goes one step further with the launch of its Typeface Arc technology. The vision behind Typeface Arc is to permit a user to specify a high-level marketing goal after which have the system robotically plan and generate all assets. These assets could include emails, images and notifications, all connected together to deliver a real cohesive customer engagement story arc.
“It's not that exciting to sit down there, type prompt after prompt, and generate a picture or a paragraph of text,” Abhay Parasnis, CEO of Typeface, told VentureBeat. “The novelty wears off pretty quickly.”
The holy grail of promoting is user personalization
Parasnis argued that lots of today's AI systems are limited to unrelated output. While generating a single piece of latest content provides moderate value, the flexibility to choreograph personalized customer journeys from start to complete remains to be essential.
“This has at all times been the holy grail of promoting,” he said.
Typeface Arc is an try and achieve this goal through the use of an integrated approach that goes beyond prompt engineering's basic piecemeal approach to generating individual pieces of content. Arc's goal is to automate the creation of personalized marketing campaigns and content by acting on behalf of the user based on their business goals.
“You just tell it what your goal is and it’ll literally create your complete marketing journey and campaign over many moments,” Parasnis said. “It generates the primary email to be sent, then a push notification that goes out a number of days later, then one other reminder email is created and every of those emails is directly targeted to the audience and product tailored.”
Using AI agents to power the following generation of AI marketing
The Typeface platform already trains and learns what an organization's brand voice and visual styles are all about.
A central component of the corporate's platform is the Typeface Graph, which provides a multimodal knowledge graph of a corporation's content. This fine-tuning and customization of Gen AI advantages Typeface Arc to drive a comprehensive marketing journey.
“We have a really strong and well-defined proprietary system where a marketer can easily describe at the best level what their goal is,” Parasnis said. “The system internally leverages our prompt corpus and Typeface Graph to essentially create a really sophisticated marketing plan.”
Parasnis explained that Typeface Arc will be considered an agent technology. Using an agent, the AI system executes workflows on behalf of users by understanding the choreography of prompts and constructing context across multiple prompts.
“We essentially created a customized planner for marketing,” he explained. “We then had to coach the LLM system to grasp not only individual prompts, but additionally the prompt choreography, similar to which prompt got here before and which prompt comes after.”
Will A/B testing be killed by generational AI-powered customer journeys?
A standard aspect of many marketing campaigns is the concept of A/B testing, which involves testing different variations to see which performs higher.
Parasnis argued that with a technology like Typeface Arc, A/B testing may grow to be a thing of the past. He explained that the technology has an analytics component that consistently evaluates the performance of the content. The system can adapt in real time to optimize messaging and content.
“I imagine A/B testing will probably be dead,” Parasnis said. “I believe the world of those marketing-oriented AI agents like Typeface Arc will construct more of a closed-loop system.”
