- Welcome, generative AI. Messagepoint is integrating generative AI (ChatGPT and GPT-4) into its platform.
- Assisted authoring upgrade. This integration will enhance Messagepoint’s AI-powered Assisted Authoring capabilities with content rewrite suggestions.
- Changing content game. The integration enables MARCIE, Messagepoint’s proprietary AI engine, to provide specific content rewrite suggestions, delivering more value to users in the process of content creation and management.
Messagepoint, a provider of SaaS-based content management for customer communications, is updating its platform by integrating it with generative AI from OpenAI's ChatGPT and GPT-4). Slated for release on May 31, the update will introduce a feature that offers content rewrites aimed at optimizing reading levels, sentiment, and content length.
The enhanced Assisted Authoring capabilities powered by OpenAI's generative AI will be integrated into Messagepoint’s MARCIE (Messagepoint Advanced Rationalization and Content Intelligence Engine). This integration means that instead of merely flagging anomalies in communication content, MARCIE will be able to offer specific content rewrite suggestions. The update is designed to support models from various vendors.
Atif Khan, VP of Messagepoint’s AI & Data Science, joined CMSWire for a Q&A to discuss the news. In this exclusive interview with CMSWire, Khan gives more details on what capabilities this update will bring and how it will capitalize on its integration of ChatGPT.
Integrating Generative AI for Enhanced Customer Communications Management
CMSWire: How do you describe what your company offers and who it does it for?
Khan: Messagepoint provides an intelligent SaaS-based content hub, which is used by marketing and customer service teams at leading financial services, insurance, healthcare and service organizations to create and manage highly personalized customer communications. The new integration with ChatGPT and GPT-4 augments the platform’s AI-powered Assisted Authoring capabilities by adding suggested content rewrites for optimizing reading levels, sentiment and content length.
CMSWire: Can you provide more insight into the specific ways that the integration of ChatGPT and GPT-4 enhances the Assisted Authoring capabilities of Messagepoint?
Khan: Messagepoint has successfully integrated the latest OpenAI-powered features into its Assisted Authoring capability, effectively enhancing its functionality and usability. Formerly, Messagepoint’s Assisted Authoring concentrated solely on evaluating the quality and integrity of communication content according to reading levels, sentiment and brand profiles. However, once any anomalies were detected, manual intervention from the user was required for the necessary corrections.
Our new release significantly improves this process. With the integration of the new generative capabilities, MARCIE (Messagepoint Advanced Rationalization and Content Intelligence Engine) can now offer specific content rewrites for a variety of applications.
Related Article: Generative AI: Opportunities and Challenges for Marketing
CMSWire: Could you elaborate on how Messagepoint’s AI platform, MARCIE, complements the AI-Assisted Authoring in Messagepoint and Rationalizer?
Khan: MARCIE is a specialized, standalone AI platform specifically designed for seamless integration into a variety of services and applications. It provides the fundamental capabilities necessary to harness high-performance AI for resolving intricate CCM use cases.
One of the defining features of MARCIE is its out-of-band processing. This unique attribute enables both internal and third-party applications to conveniently integrate and customize content empowered by MARCIE, fostering significant improvements in content quality. With this attribute, content improvements can be done both just-in-time and just-in-place, making MARCIE a reliable and flexible solution for complex integrations such as Messagepoint’s Assisted Authoring tool.
CMSWire: Why did Messagepoint choose to integrate with ChatGPT and GPT-4 specifically? Are you considering other large language model solutions for future integrations?
Khan: Our choice to partner with OpenAI was influenced by several key considerations. Paramount among these was their robust data privacy and usage policies, aligning with our commitment to secure and responsible data handling. Additionally, their swift pace of innovation provides us with a dynamic and forward-thinking platform. Lastly, the flexibility to customize our offerings to meet our clients' unique needs was a critical factor in our decision-making process. Our collaboration with OpenAI underpins our strategy to deliver customized, secure and innovative solutions.
My conviction is that we will have diverse types and variations of large language models available to us. Anticipating this, we have designed our solution to support models from various vendors. Our primary motivation is to ensure we align the most suitable tool with each customer's specific needs and challenges. Our solution's flexibility and adaptability position us strongly for the evolving landscape of large language models.
Related Article: Generative AI Solutions for the Contact Center
Evaluating Reading Levels, Sentiment and Length
CMSWire: How does the AI-powered content generation take into account desired reading levels, sentiment and length?
Khan: Content dimensions such as size, reading level and sentiment represent crucial facets of raw text that aid in evaluating key characteristics, including content health, efficacy, impact and user engagement. With the ability to precisely measure these parameters, we have achieved a substantial comprehension of content differentiation based on these attributes. This understanding fuels our content generation process, playing a pivotal role in the pre-processing, prompt-engineering and post-processing stages of our generation pipeline.
For example, when asking the machine to rewrite text to have a better sentiment, we ensure preservation of meaning, key entities and topics and, most importantly other, CCM-related concepts.
Making Customer Communications Legible
CMSWire: How does your technology help to address the challenge of industry jargon and legalese in customer communications? Can you provide some specific examples?
Khan: Often, text riddled with industry-specific terminologies and complex legal phrases poses a comprehension challenge, increasing the readability burden. Our MARCIE AI platform provides an array of readability detection algorithms. These tools can be deployed to effectively gauge the readability impact of difficult-to-comprehend text. With our generative AI solution, MARCIE can now proactively offer suggestions to simplify the text, enhancing its ease of reading.
Here is an illustrative example of how MARCIE can simplify a complex sentence:
Original: "Party A shall indemnify Party B against all claims, damages, etc."
Rewritten: "Party A will protect Party B from any problems, harm and so on."
In addition to improving readability, our initial offering also focuses on summarizing text and discerning sentiment, thereby delivering multifaceted benefits.
Making AI-Generated Content Compliant
CMSWire: How do you ensure that the AI-generated content is compliant with regulations in different sectors?
Khan: Today, compliance requirements remain the responsibility of the content authors. That said, the Messagepoint platform leverages a modular approach to content management, which enables a content object containing text or an image to be centrally managed and shared across multiple touchpoints and channels. This capability, called SmartText, is often used by customers to house regulatory language, such as disclosures, that appear in multiple places. The generative process will respect the existence of a SmartText object and not overwrite it, preserving that disclosure requirement. If a user wants to optimize that disclosure content, it can be done by evaluating that specific message.
CMSWire: What would a user of this software actually see when they use it?
Khan: Messagepoint’s Assisted Authoring capability automatically detects and flags readability, sentiment and brand issues in the content through MARCIE. Upon seeing a red or yellow indicator that identifies an issue, users can request a suggested rewrite through the Assisted Authoring panel. MARCIE will submit the content to ChatGPT and protect any variables or shared content objects. Upon receiving the suggestion, users can click to accept it, confident that any content formatting will be preserved.
CMSWire: What are the price points for new and existing customers?
Khan: Assisted Authoring is available as an add-on module to the Messagepoint platform, which is sold on a subscription basis.
CMSWire: What other ways does Messagepoint plan to leverage AI to enhance its services in the future?
Khan: We are steadfast in our belief that the customer communication management industry is poised to undergo a transformative shift driven by artificial intelligence. We anticipate content of all types, sizes, formats, languages and modalities to traverse both existing and new delivery touchpoints. Messagepoint’s AI strategy is intensively centered on comprehending content from any source and enriching it within the context of its use.
Have a tip to share with our editorial team? Drop us a line: