Magid’s AI experts discuss how newsrooms can “uplevel” their journalism quality with Collaborator’s Analyze

Magid’s AI experts discuss how newsrooms can “uplevel” their journalism quality with Collaborator’s Analyze

Alberto Melgoza, Magid’s Chief Technology Officer and Head of Products, and Steph Smelewski, Product Manager for AI Applications, discuss Magid’s Analyze tool, which is designed to harness the power of AI to help journalists radically improve the quality of their content.

What’s the value of the Analyze tool? Why did you build it?

AM: Ultimately our vision for Collaborator Newsroom is to be the most helpful AI companion or “collaborator” to the journalist and solve the problem of cross-platform versioning. That module is the one that we call Translate. We’re seeing it already have transformational impacts across newsrooms. 

Another area that we’re seeing be transformational within the newsroom is AI that helps journalists  uplevel the journalistic quality of the stories that newsrooms are producing.

That’s where we found that the Collaborator Newsroom Analyze module can be really impactful and transformational. And that’s why we built it: to help a newsroom really uplevel the quality of the content that they’re creating. 

How is it helping journalists “uplevel” the quality of their journalism?

AM: We leaned on the fact that Magid has spent seven decades working every day inside newsrooms. We’ve worked with thousands of journalists day in and day out, and have developed extensive knowledge, experience, and a library of journalistic best practices. We have come to really understand and identify what makes good journalistic content that resonates with audiences and helps newsrooms reach their goals and reach their audiences.

So we took all of that knowledge and trained the Collaborator Newsroom Analyze module to become the realtime companion to the journalist. It’s like having all of that Magid knowledge in your pocket as you create your story. Analyze is an evaluation layer that helps you evaluate your content across the dimensions that we know are most critical for optimal journalistic content.

What are those dimensions?

SS: So, there are three main score categories. We have a Balance and Fairness score, Quality score, and Spelling and Grammar score. Each category has dimensions, through which Analyze helps journalists get a full picture of what it is they’re writing and evaluates how effective it’s going to be for their audience.

Within the Balance and Fairness score, we have four dimensions:

  1. Facts and Opinions evaluates how well factual statements are attributed and distinguishes them from opinion-based statements. 
  2. Use of Language focuses on the presence of loaded or biased adjectives that may affect neutrality of the story. 
  3. Transparency measures whether the story communicates unanswered questions and points out any missing information about process.
  4. Voice Representation analyzes the diversity and the balance of perspectives within the story. 

Then we have our Quality score, which looks at Writing Quality, Structure and Organization, and Language.  

  1. Writing Quality focuses on clarity, consistency, and adherence to optimal journalistic style across different formats.
  2. Structure and Organization determines if the story follows a logical flow. Does it have a clear beginning, middle, and end? Is that clear and communicated to the reader as they’re reading the story?
  3. Language identifies if there are any cliches, jargon, or other language that doesn’t align with the Magid quality standards and that can affect readability. 

Then we have Spelling and Grammar, and Syntax. 

  1. Spelling and Grammar looks for grammar and punctuation errors that impair readability.
  2. Syntax looks at things like tense consistency, run-on sentences, and stylistic choices that could negatively affect readability. 

Grading system for Collaborator Analyze tool

How does Analyze “score” each story?

SS: Within our Analyze dimension system, we have a color scoring system. 

  1. Dark green tells the journalist this story is good to go and we have no further recommendations per the Magid standard guidelines.
  2. A light green rating indicates there is some minor advice or suggestion for improvement given, but overall the story is good to go.
  3. Yellow means that revisions are advised before publish. Something is conflicting with our Magid best practices and we offer direct feedback to improve it.
  4. Finally, a red score indicates that something must be revised before publish. This could be something like clearly biased language, major typos, or glaring grammar or syntax issues.

AM: And regarding the coloring, we built Analyze to be non-judgmental. To be analogous to an X-ray for doctors. An X-ray doesn’t tell the doctor what the condition is, what to prescribe, what to tell the patient, or what to do next. It simply presents a picture. And that’s how we built Analyze. It’s that evaluation tool that gives the journalist a clear picture of what they’ve presented. The journalist is always in the driver’s seat and decides what to do next with that information.

Can we see it in action?

Steph Smelewski walks us through Collaborator’s Analyze tool.

From a technical standpoint, what’s special about Analyze?

AM: Yeah, let’s talk about how this works and how we built the Analyze module and the complexities that go under the hood that have been abstracted from the journalist. This module fits in really well with Collaborator’s AI enterprise architecture.

We built Collaborator from the ground up to be an enterprise AI system where all that expertise and knowledge that we’ve been talking about is part of the AI knowledge base. It also lives with the individual newsroom’s brand voice and brand differentiators, and any additional consumer audience data and insights. All of those live together in the AI knowledge base, within a general architecture that has now become the most effective architectural paradigm when building AI applications.

So these components are the enablers, or the building blocks, for Collaborator’s governance evaluation layer when you’re building AI applications. There’s a lot of discussion and very important work going into ensuring trust and accuracy for AI applications. In fact, we at Magid have exciting features coming out very soon that further ensure trust and accuracy with our AI tools.

Within Collaborator, you have this evaluation layer that knows and cares about and looks for journalistic best practices being applied to your content. These are all enabled by Magid’s enterprise AI system architecture. 

Where are newsrooms seeing value in Analyze?

AM: Here’s where we’re seeing Collaborator’s Analyze having the biggest impact.

First of all, journalists get that real-time feedback as they create content, which helps them create the best quality content possible. We’re already seeing that, on average, half of the best digital content for a newsroom is content that has been created with Collaborator, meaning stories earning the most traffic and achieving other key newsroom KPIs.

Another transformational impact that we’re seeing Analyze have in newsrooms is that it’s being integrated into editorial workflows. Analyze is really very naturally integrated into these existing newsroom workflows as an accelerator for the review processes. Again, it’s like having a Magid expert review every piece of content that flows through your newsroom. 

It’s also effective as a training tool. That’s because Analyze has been built with well-established journalistic principles and best practices. And because of the nature of its real-time feedback, it’s becoming widely used to train incoming journalists on where the quality bar is.  

Finally, my favorite example of a newsroom use case, and the most interesting one to me, is how Analyze is helping the newsroom define and track content quality KPIs.

As I mentioned earlier, a key value proposition of Collaborator is helping a newsroom uplevel the quality of its content. For a newsroom to truly know if that is happening, it’s important to be able to measure that through KPIs. And the Analyze scorecards we distribute to newsrooms have become that way to measure and track how the newsroom is evolving and upleveling its content quality as users get this feedback. And we’re already seeing how that quality continues to go up as the tool is being deployed and adopted in the content creation workflows.

Collaborator Newsrooms is designed to supercharge your newsroom’s efficiency and quality, while ensuring top security and accuracy. Get more competitive using cutting-edge AI technology that empowers the journalist in the driver’s seat. Leverage Magid’s 7 decades of journalism expertise at your fingertips every time you create content. Get a personalized demo and learn more here.