Most communications leaders I talk with have a version of the same problem. They understand that something has changed. Buyers are arriving to sales conversations already formed in their opinions. Website traffic is declining. Coverage that used to drive pipeline now seems to disappear into a void. They suspect AI has something to do with it, but no one in their organization can explain exactly what to do about it.
The category of "computational communications" is my attempt to name what is actually happening, and to describe the work that follows from it.
Computational communications is the discipline of designing how an organization is understood, evaluated, trusted, cited and recommended across human and machine-mediated environments. The output, whether a press release, an article, a keyword ranking, or a citation, is not the point. Understanding is.
This piece offers a definition, explains why the shift matters now, and closes with a working glossary of the terms you will encounter as this discipline develops.
What Changed, and Why It Matters Now
Corporate communications was designed for a human-mediated information environment. Journalists interpreted companies. Search engines indexed their pages. Analysts evaluated their products. Buyers visited their websites, read the coverage, and formed views through a relatively predictable sequence of human touchpoints.
That sequence has been disrupted. Not replaced, but significantly altered.
Forrester's 2026 State of Business Buying research, drawn from its Buyers' Journey Survey of nearly 18,000 global business buyers, found that 94% report using AI during their buying process. In the same survey, buyers named generative AI or conversational search as a more important information source than vendor websites, product experts, or sales representatives. Separately, G2's Answer Economy report found that 51% of B2B software buyers now start their research with an AI chatbot more often than with Google, up from 29% in G2's 2025 report.
The implication is straightforward, even if the response is not: a large share of your market is forming a view of your company before a person on your team is aware they exist.
That is a communications architecture problem, not a search problem.
AI systems do not simply distribute information. They make probabilistic choices about which sources to retrieve, which claims have enough corroboration to be treated as true, which company belongs in a category answer, and which vendor deserves a recommendation. The G2 data shows how consequential those choices are: in its March 2026 survey of 1,076 B2B software buyers, 69% said they chose a different vendor than they had originally planned because it was part of a chatbot's recommendation, and one in three purchased from a vendor they had never previously heard of.
Those outcomes are not random. They reflect the quality, structure, consistency, and third-party corroboration of the information environment a company has built, or failed to build, around itself.
Traditional PR earns coverage. Traditional SEO earns rankings. Computational communications designs the information environment from which the answer, and ultimately the decision, is formed.
How Computational Communications Differs from Traditional PR and Marketing
The question I get most often is some version of: "Isn't this just PR plus SEO plus some AI optimization?" The honest answer is no, and the distinction matters.
Traditional PR, digital marketing, AEO, and GEO are each optimized for a specific channel or output. PR earns media coverage. SEO earns search rankings. AEO (answer engine optimization) helps a company appear in a specific AI answer. GEO (generative engine optimization) focuses on how content is retrieved by generative models. Each of these has value. None of them, on their own, addresses the underlying problem.
The underlying problem is that your company's reputation is now a computational output: a probabilistic synthesis generated across many signals, sources, and surfaces. No single channel intervention changes that synthesis.
The Difference in Practice
| Discipline | Primary objective | Unit of work |
|---|---|---|
| Traditional PR | Earn media coverage | The placement |
| SEO | Rank in search results | The keyword |
| AEO / GEO | Appear in AI answers | The citation or prompt |
| Computational Communications | Design how the organization is understood | Understanding |
The distinction is not cosmetic. A company can earn significant coverage and still be poorly represented in AI answers, because the sources AI systems rely on may be outdated, contradictory, or absent. A company can rank well in organic search and still be invisible in the AI-generated summaries that now appear above those results. Third-party trackers estimate that AI Overviews appear on roughly 43% to 48% of Google queries, and Ahrefs found that only 38% of pages cited in AI Overviews rank in the top 10 for the same query.
Computational communications asks a different question at the start: what does the market actually understand about us, across every surface where understanding forms?
What the Work Actually Involves
The answer to that question produces a different kind of program. It includes the elements traditional communications has always required: narrative, messaging, earned media, thought leadership, and stakeholder engagement. But it adds layers that most communications programs do not currently include:
- Structured knowledge and entity design: Making an organization's identity, expertise, products, and relationships legible to machines, not just readers.
- Evidence architecture: Building independent, third-party corroboration behind every priority claim, so that AI systems retrieving information about the company find consistent, credible support.
- Content supply chain: A continuous production system that keeps the information environment current, because AI systems are not reading last year's press release.
- Measurement of understanding: Tracking not just coverage and rankings, but whether the market, including AI systems, actually grasps who you are and why you matter.
This is explored in more depth in The Age of Agentic Awareness, which examines how autonomous systems now participate in forming and distributing organizational narratives.
A Point of View
I want to be direct about where Actual Agency stands on this, because there is a lot of noise in the market right now and I think clarity is more useful than positioning.
We believe the communications function is at an inflection point. Not because AI is new technology to bolt onto existing workflows, but because the environment in which communications operates has structurally changed. The information landscape is now computational. The tools and methods built for a human-mediated environment are necessary but no longer sufficient.
We also believe that most of the current industry conversation about "AI in PR" is pointed in the wrong direction. It focuses on production: using AI to write more pitches, generate more content, automate more outreach. That misses the more important question, which is whether the information environment around your organization is actually producing the understanding you need.
The question worth asking is whether the output is actually producing understanding, not just volume.
A press release that earns no coverage and is never retrieved by an AI system did not communicate anything. A thought leadership article that builds genuine authority in a category, gets cited by third parties, and consistently appears when buyers ask AI systems about that category, that is communications working.
We see this problem repeatedly in B2B technology companies. They have strong communications programs by traditional measures. Good coverage. Consistent content. Active social presence. And yet when you ask ChatGPT or Perplexity to compare their category, the company is either absent or misrepresented. The information environment has not caught up to the strategy.
Computational communications is our answer to that gap. It is not a replacement for what good communications teams already do. It is the next operating layer, built for the environment those teams are now working in.
You can read more about how we think about this work on our beliefs page and in our overview of executive thought leadership for AI search.
A Working Glossary
The vocabulary around this space is unsettled. Terms get used interchangeably, defined inconsistently, or invented for marketing purposes. What follows is how we use these terms, with the reasoning behind each definition.
Computational Communications
The discipline of designing how an organization is understood, evaluated, trusted, cited and recommended across human and machine-mediated environments. The goal is not coverage or rankings in isolation, but whether the market, including AI systems, accurately grasps who you are, what you know, and why that matters to the decision in front of it. Earned media, SEO, AEO, and GEO are capabilities within this system, not substitutes for it.
Agentic Awareness
An organization's ability to understand how autonomous and semi-autonomous AI systems participate in the discovery and explanation of its narrative. AI systems make probabilistic choices about which sources to surface, which claims to treat as credible, and which companies to include in category answers. Agentic awareness means understanding those choices and designing for them. See The Age of Agentic Awareness.
Answer Engine Optimization (AEO)
Structuring content so that AI answer engines, including ChatGPT, Perplexity, and Google AI Overviews, retrieve and cite it accurately. AEO is one capability within a broader computational communications program, not the program itself.
Generative Engine Optimization (GEO)
A related discipline to AEO, focused specifically on how content performs within generative AI systems. Where traditional SEO targets search rankings, GEO targets how generative models retrieve, summarize, and attribute content. Research published on arXiv has examined how content characteristics, including statistics, citations, and authoritative sourcing, influence visibility in generative AI outputs.
Entity Design
The process of making an organization's identity legible to machines. This includes structured data, schema markup, knowledge graph entries, consistent naming conventions, and the relationships between an organization's people, products, claims, and sources. A machine can only retrieve what it can parse. Entity design ensures the organization's core information is structured in a form machines can read and connect.
Evidence Architecture
The deliberate construction of independent, third-party corroboration behind an organization's priority claims. AI systems weight claims that are supported by multiple credible, non-self-referential sources. Evidence architecture is the work of building that support: earned media, research, analyst coverage, customer validation, and expert attribution that corroborates what the organization says about itself.
Narrative Architecture
The structural design of an organization's communications across all surfaces. Narrative architecture is not a messaging document. It is the system of claims, evidence, entities, and relationships that, taken together, produce a coherent and consistent picture of the organization wherever it is encountered, whether by a journalist, a buyer, or a large language model.
AI Visibility
A measure of how frequently and accurately an organization appears in AI-generated answers to relevant queries. AI visibility is distinct from search visibility. A company can rank well in organic search and be largely absent from AI answers, because the citation pools used by systems like ChatGPT and Perplexity overlap only partially with Google's top-10 results.
Information Environment
The totality of sources, signals, and content that AI systems, search engines, journalists, analysts, and buyers encounter when researching an organization. The information environment includes owned content, earned media, third-party reviews, social content, regulatory filings, analyst reports, and any other publicly accessible material. Computational communications treats the information environment as the primary object of design and management.
Machine-Mediated Environment
Any context in which an AI system, algorithm, or automated process participates in how information about an organization is retrieved, summarized, compared, or recommended. The machine-mediated environment now includes AI answer engines, search algorithms, LLM training data, retrieval-augmented generation systems, and AI-powered research tools used by buyers, journalists, and analysts.
Definitions are useful, but they are not the work. The work is building an information environment that accurately and consistently represents what your organization knows, what it does, and why it matters, across every surface where a buyer, journalist, analyst, or AI system might encounter it.
That is a different kind of communications program than most organizations currently run. It requires different skills, different tools, and a different way of measuring success. It does not require abandoning what good communications teams already do well.
The discipline is still developing. The vocabulary will continue to shift. Some of what we call computational communications today will simply be called communications in five years, because the environment it describes will be the only environment there is.
That is probably the most important thing to understand about this moment.
Frequently Asked Questions
- What is computational communications?
- Computational communications is the discipline of designing how an organization is understood across human and machine-mediated environments. It focuses on the full information environment around a company, including owned content, earned media, citations, search, and AI-generated answers.
- How is computational communications different from traditional PR?
- Traditional PR is built to earn coverage through human intermediaries. Computational communications goes further by shaping how a company is retrieved, summarized, compared, and recommended by AI systems, search engines, and buyers. Coverage still matters, but it is no longer the whole job.
- How is computational communications different from marketing?
- Marketing usually focuses on demand generation, campaign performance, and owned-channel conversion. Computational communications focuses on the integrity and legibility of the broader information environment, so the organization is correctly understood before a buyer ever reaches a campaign asset.
- Why does computational communications matter now?
- Because buyers increasingly research vendors through AI-mediated surfaces before they speak with sales. That changes where reputation is formed. If the information environment is inconsistent or stale, the market may form a poor view of the company before a human team can correct it.
- What does Actual Agency mean by agentic awareness?
- Agentic awareness is the ability to understand how autonomous and semi-autonomous AI systems participate in the discovery and explanation of a company's narrative. It means knowing what AI systems are saying about you, which sources they rely on, and where the representation diverges from reality.