AI in PR – opportunities and risks

AI is an analytical tool

At its core, AI tools are analytical technologies. To deliver, they consider the request, look at data, reason using the available data, and provide a response. Therefore, if for a moment we ignore the specific AI model’s capabilities (which in general you need to consider in your AI strategy), the AI deliverables depend on the data it works with. Hence, the critical importance of organizational data.

Internal vs. external data

From the point of view of the organization, there are two kinds of data – internal and external. Simply put, internal data is the data generated in the context of the organization’s operations. These data are owned and stored by the organization.

External data is all data, unrelated to the organization’s operations. The data used to build the AI models is public data – data accessible to anyone. Imagine websites and Wikipedia.

In the PR context, internal data is our media contacts database and our interaction with media, while external data is the media representatives’ publications and content, social media profiles, and social media engagement, unrelated to our interactions.

In a perfect world, the organization should store as much as possible internal data, enrich it with external data for greater context, and deploy AI for fast and high quality analysis and interaction with the resulting unique data.

These data are unique, because of the internal data. While external data is available to anyone, the internal one is proprietary and unique to every PR team. Internal data is critical for the PR performance and competitiveness of the organization.

Regrettably, the internal data management has been the most common and major organizational problem since the beginning of enterprise technology. Usually, the smaller organizations don’t keep records on their operations, which limits their ability to grow and learn and develop their capabilities.

Subsequently, I noticed that when considering AI, marketing and PR professionals mostly focus on the external data in the form of generative AI for creating press content or media monitoring applications.

There is already enough research today that shows that AI relying solely on external/common data is not “original”. For example, such a tool cannot propose sound and original PR and marketing strategy.

What is AI context?

The integration of external and internal data provides AI with context. Deploying the AI technologies for competitive advantage is highly dependent on the AI context.

So what is AI context?

“AI context refers to the structured information, memory, and situational data that AI systems use to generate relevant, accurate responses. This includes the immediate prompt, prior interactions, user intent, and any connected data sources that help the system “understand” what’s happening.” Salesforce.

AI context is more the “what’s happening” understanding. As it is related to intent, more importantly it is about “why have been asked this” and what “my limits and reasoning scope are”. As we can see from the above definition, AI context relies on two major aspects – interaction memory and data sources.

Firstly, it needs to have memory to put every subsequent request in the context of the human’s interaction. This means that the agent considers all previous interactions. For example, the user does not need to specify data sources, press release styles, or steps along a process. The AI will continue where it left off, just like a human teammate.

Secondly, the AI’s situational awareness and relevancy of response is determined by the data it works with. This is directly related to the external and internal data discussed above. The most simple example is creating PR and marketing deliverables, subject to brand rules, including tone of voice.

The internal-external data relationship is bi-directional. AI could be grounded with internal data, but also could be enriched with external data. For example, in PR outreach you have certain outcomes – a journalist must have opened, not opened email, published or not published a pitch. If limited to these data, the AI would not be able to recommend best course of action. But adding external data for the media and journalist – their history of publications, their brand sentiment, or their frequency of publications, role in the organization and more, the AI can reason about “why” the particular outreach outcomes. In the context of the PR process, the AI can reason and propose the most relevant addressees for this particular pitch.

In this way, we enrich our understanding by combining behavioral with psychographic and demographic data.

In Nippy, for example, the AI MRM Search creates a semantic order of the most relevant journalists, considering all available data. This order determines your best chance for media pitch success. And as it is semantic, you don’t need to use general categories such as “mobile phones” or “consumer electronics”. You can just type the brand’s name – Samsung or Google Pixel and the AI search will understand your intent, because it will understand your intent and related with journalists that have written about mobile phones, even if not tagged in your MRM.

In conclusion, AI context is the most critical for the AI ROI and process’ efficiency AI characteristic.

PR applications of Artificial Intelligence

From point of view of the organization, we can divide AI in two broad categories – content generation and process facilitation.

Generative AI for content creation

Most professionals implicitly consider generative AI when they think about PR AI. Its benefits to the PR and marketing operations are apparent – greater speed, quality, and volume of content creation, thus greatly improving the organizational efficiency.

But what about effectiveness and expected outcomes? The PR professional is still responsible for the PR message. They need to be clear about this as the AI won’t be able to deliver the expected content.

Gen AI could save a lot of time in researching, drafting, styling and adding external data context to a press release. But when all PRs have access to the same model that use same external data, where is the competitive differentiation in the output? I guess, you could work with the AI agent’s instructions, but this is it.

And have you heard about AI slop?

“AI slop is a derogatory term for low-quality, mass-produced digital content created using generative artificial intelligence with little to no human editing, curation, or creative effort.” Wikipedia.

I think that the biggest contributor to AI slop is lack of context, not lack of human interaction.

A press material AI agent can bring value only if the PR professional already has sound understanding of their PR strategy, target media and message. Furthermore, they need to be able to work with AI agents to create content that is engaging, to the message and in the desired style. This would be challenging for an inexperienced PR professional.

In conclusion, gen AI can improve efficiency in PR content creation, but will be limited in its contribution if not grounded in interaction history and internal and external data.

PR operations AI

What usually is missing in the discussion is the AI in support of the PR operations.

There are two major processes in the PR organization: building and sustaining media relationships and creating and managing integrated marketing and PR communications campaigns. Both have their challenges and specifics. Building sustainable media relationships require personal contact, while campaigns execution is subject to internal rules and processes, more than anything else.

One of the biggest problems still is PR execution. Recent research shows that despite of all technologies, around 85% of the campaigns are with missed launch date. Obviously, this negatively affects both effectiveness and efficiency of the PR operations. The major reason for this are internal processes that no AI can remedy for.

Similar for personal contacts and media relationships. No AI can supplant this.

But deploying AI to the internal data, enriched with external context, can greatly improve the PR process.

The media relationship management requires media database and context. The media you have in your Nippy and the data on them are competitive advantages. AI is great in this case. For example, in Nippy you can analyze the editorial topics for every media contact, do AI search for media according to key topics and other parameters, and create segments for better targeting.

And you could do this with a simple natural language request: “Create an email pit to business journalists in Ireland, who has written about data centers. Include subject and email copy for a press release pitch about…”

Using AI can greatly improve the quality of media outreach and speed up the process, improving both efficiency and effectiveness.

Operational AI is great for analysis

To evaluate the effectiveness and efficiency of the PR process you need to analyze many metrics along the process. Now, with AI agents you can converse, ask questions and extract insights. They can even create charts for you. And they can do this in context, using both internal and external data. And, they will create you the charts you need for your reports.

How PR organizations should deploy AI?

For AI to bring value to your organization and PR process, you need to consider the user cases for implementation, the integration of internal and external data, as well as employee training and skills.

Consider PR process use cases

The first step is to determine where in your processes AI will bring greatest value. For that, you need to consider the following questions.

To what aspect of your PR operations you need to bring value to? Is it client relationship management or internal efficiency.

How do you create value? How do you define success and high performance in your organization? Is it the number of publications or the quick project turnaround?

Where is your competitive advantage? What services are core to your business model?

For example, from the beginning in our agency we have decided that core to our business model and differentiation is strategy, messaging, writing, and content creation. Speeches, press materials, any content. We have never outsourced this externally or delegated this to junior staff.

Some PR use cases:

1. Content creation
   • Press releases
   • Speeches
   • Emails
   • Reports
   • Charts
2. Process management
   • App interaction – use conversations to execute, in addition to traditional clicks interaction
   • Media segmentation
   • Media enrichment
   • Research
   • External execution – online forms, bookings
   • Alerts

• Analysis
  • Key topics analysis
  • Sentiment analysis
  • Behavioral analysis
  • Performance analysis
  • Operations analysis – outreach effectiveness, PR performance, media response, team performance, any types of analysis

Map your processes, identify critical junctures and stages to improve or strengthen.

Connect internal and external data

Record every possible interaction along the PR process. Today, these usually are digital interactions that don’t require significant administrative effort.

For example, every sent outreach and resulting outcomes – sent invitation, delivered, opened, clicked, spam, confirmed, attended event must be saved in your PR system for further analysis.

Connect these data with every media contact and enrich it with as much as possible external data. For example, in Nippy you have all interactions with every media representative AI searchable together with their publications, AI extracted key topics from their publications, the information in media monitoring, even the notes you have made in their profile.

Train young professionals without AI

Create a development program for young professionals without extensive AI usage. Might sound counterintuitive, but young PR professionals need to develop two core skills: writing and feeling for media.

Both are dependent on reading. The need to read media and written content as much as possible. Thus, they will develop instincts and understanding about what messages work with media and how media write and adapt corporate messages.

Understanding how to express in writing is dependent on two things – reading and writing. Young professionals need to write daily all types of content. When they show that they can create written content, they could start using AI tools to enhance their content. More importantly, they could start checking and proofing AI content.

Conclusion

While AI brings great benefits for the PR organization, as any other tool it needs to be implemented with clear understanding of the strategic direction of the department, agency, or team in such a way that AI can enhance team’s strength, improve the PR process efficiency, and bring value to internal and external stakeholders.

PR organizations need to start accumulating internal data, enrich it with external data, and implement AI that is grounded in both and strategically pointed to the aspects of your PR process, which you find strategically most important for the competitiveness of your PR function.