Interest still ahead of adoption
If you manage your organization's strategic communication function, you and your team are likely wrestling with how to best leverage AI. Everyone is interested in AI, yet there seem to be relatively few success stories. BCG's survey of more than 200 corporate affairs and communications leaders in the United States and United Kingdom further highlights the AI problem. Sixty-eight percent described their function as an AI laggard. Only one in ten had systematically integrated AI into workflows, while 35 percent cited a lack of operating-model design capability as a leading obstacle.[1]
Corporate affairs leaders are trying to adopt AI while their functions are already changing. Deloitte's 2026 study of 42 corporate-affairs leaders found that 83 percent of their functions were undergoing significant change and 66 percent were changing their operating models. But only 24 percent had a formal AI strategy. Most teams were still exploring and experimenting, with content creation and summarization among the most common uses.[2]
A broader Oxford–GlobeScan survey of 294 senior professionals found the same pressure to change. Seventy-four percent said Corporate Affairs needed some form of revision. Respondents most often called for clearer strategy, focus and priorities (35 percent) and changes to operating models, governance and internal alignment (30 percent). They also identified impact measurement, data and AI capabilities as areas that need work.[3]
The studies are not interchangeable: their samples vary, and BCG combines corporate affairs with communications. The value they provide is context: they show that in the environment in which strategic communication lives AI interest is high, but successful integration is low.
The answer isn't devoting AI resources to disconnected discrete tasks such as drafting, summarization, monitoring and research. It isn't access to a better model. It is integrating AI into a structure that leverages its strengths while mitigating its weaknesses.
AI isn't a panacea
Most people's exposure to AI is limited to LLM chatbots: open a browser, navigate to ChatGPT, Claude or Gemini, and ask a question. This prompt-response model is useful in many contexts. But anyone who has used these tools knows that they can confidently deliver incorrect information and be sycophantic; concerns also remain about privacy and data use by frontier labs.
Over the past year, however, AI has increasingly taken the form of agents. Users will never see some agents, which toil away in the background at tasks given to them by an application that controls them. Others can be created and controlled directly: you may recall OpenClaw, which began in late 2025 and prompted some users to buy Mac Minis to run personalized agents that manage email inboxes and calendars.[4] More advanced users working in OpenAI's Codex, Anthropic's Claude Code and Cowork or in Cursor are using complex multi-agent systems to create applications or automate workflows.
While AI chatbots provide direct conversational access with an LLM via a prompt in which users can ask whatever they like, an AI agent is essentially a well-crafted, bounded, goal-directed prompt in which the LLM can choose and use tools, observe the results, maintain relevant context and iteratively decide what to do next until it reaches a stopping condition. This means they can operate independently, and are certainly the more powerful manifestation of AI. They are also the most dangerous, as the recent OpenAI agentic hack of Hugging Face demonstrates.[5]
The fact is that neither AI chatbots nor agents provide comprehensive solutions to strategic communication challenges out of the box. While they can be helpful in limited ways they each have significant downsides, that if not properly accounted for, present more risk than upside.
Build the system first
We built the Strategic Communications Intelligence (SCI) platform to govern the part of an organization's strategic communication function focused on independent external stakeholders. SCI brings the information and workflows required for that work into a centralized system with distributed, accountable ownership. There's a reason we're the first to do so: it's really hard. The larger and more complex an organization is, the more information and workflows the system must connect. We made SCI feasible by defining its scope carefully and using AI agents to automate key workflows, extend the platform's capabilities and reduce the burden on users.
Strategic communication as a broader function may address internal and external stakeholders. SCI is narrower. It focuses on independent external stakeholders whose actions can affect strategic outcomes but whom the organization cannot control. Customers and suppliers ordinarily remain in the systems governing those relationships; they enter SCI only when acting in a distinct external-affairs capacity.
With a narrower set of relevant information and workflows to manage, we were able to better structure the application to provide governance of the entire suite of logical records required for a strategic communication program. Even so, the application wouldn't be feasible without the integration of AI agents to perform a number of limited but key tasks.
First, an overview of the structure we created. Built to serve as a system of record for strategic communication, we established what we call the strategy-to-stakeholder-to-outcomes chain that relies on the following links:
- Broad organizational priorities that connect to specific objectives and their desired outcomes.
- Independent external stakeholders who must take action for those objectives to be achieved and desired outcomes met.
- Communications, engagement and operations activities that impact those stakeholders.
- Talking points crafted to inform and align those stakeholders.
The platform itself establishes the chain, the logical records serve as the individual links, and we created an evidence subsystem that provides user-verified citations for each link across the chain.
The value is not in the number of fields or record types these links represent. It is that a record has a defined meaning, an accountable owner, a current status, explicit relationships to other records and evidentiary support that subject matter experts can validate and practitioners examine.
This architecture preserves distributed institutional ownership. The subject matter experts within policy, program, project or operations teams are assigned records that represent the objectives, stakeholders and operational activities that impact those stakeholders. The strategic communication practitioners are assigned records for the communication and engagement activities that target stakeholders and the talking points crafted to inform and align them. This distribution of records along practitioner/SME channels ensures that the right people update the right records, while logical centralization makes their records mutually legible. The resulting discipline ensures that users can only update records assigned to them, but all users can read all records. Drumbeat reports provide the synthesis that is sent to organizational leaders to ensure internal alignment.
This design provides ideal opportunities for AI integration. Instead of working from an undifferentiated collection of documents, it works against bounded, current, organization-specific context. It can connect a new item of evidence to the records and relationships that give it meaning. Its outputs can point back to those records and sources. The subject matter experts and practitioners closest to the links in the chain can review AI-provided recommendations, authorize updates to their records and request ad hoc queries across the entire chain.
The same structure creates a correction loop. When a record owner rejects evidence, revises an interpretation or changes a status, the system captures the accountable decision. Subsequent analysis begins from the maintained record, not from an untraceable prior answer. AI assistance therefore remains subordinate to institutional ownership while empowering it.
Scale through AI
SCI as a platform provides the structure needed to govern the strategic communication function, but that structure creates the need for maintaining many records across the organization. This is the scaling problem: the more records requiring maintenance, the more human capital required to establish and maintain them. A single record is assigned to a primary owner who is accountable for the state of the record, and any number of delegate owners who can help maintain it. Even so, each record has many data fields that need to be continually updated if the strategic communication value proposition is to hold. AI allows SCI to scale by reducing user burden in three ways: evidence discovery, contextual assistance, and analysis and synthesis.
Evidence discovery. AI agents gather and evaluate supporting material relevant to governed records by scouring internal and external information sources. Record owners closest to the material review the presented evidence and determine whether to validate it or reject it. Validated evidence is absorbed into the system and used in ongoing analyses. This helps teams keep the institutional picture current without relying entirely on manual surveillance.
Contextual assistance inside the application. Certain entry form fields include an assistance button that activates an agent that conducts analyses on other fields within the record, supporting evidence or even other records and record types. The agent then presents a proposed value for that field that the user can accept, edit or reject. Additionally, an in-app chatbot provides daily recommendations for updates to fields of records owned by that user. The chatbot can bring forward connected context and reduce navigation and assembly work, but it does not hold independent authority over the record or the decision.
Analysis and recurring synthesis. Cross-record analysis surfaces relationships, evidence currency, inconsistencies, posture changes and matters requiring attention. The insights gained from these analyses result in a number of reports that go to different audiences within the organization. The weekly Strategic Communication Brief (SCB) is created by agents that synthesize information from all relevant records and provides status updates on strategy-to-stakeholder-to-outcomes chains, flagging significant advances and risks to ensure internal alignment. The SCB is a leadership-level report that is sent to a controlled distro. An organization-wide SCB Annex is also generated weekly, but is not distributed; SCI users can review the Annex within the application, providing them with a comprehensive level of operational detail across the entire record set. Lastly, individual record reports provide a current view of a specific record, and are available to all SCI users to ensure continuity across the userbase.
How does this work in practice? Consider a wholly hypothetical utility transmission-siting matter. The regulatory team owns the filing and procedural record; government affairs owns relevant official relationships and political conditions; community relations owns local meetings and commitments; communications owns current public material; and the project team owns operating developments. A new filing, stakeholder statement, meeting note or project decision indicates that a consequential stakeholder's position has changed or that an organizational commitment is at risk.
Within SCI are records that each represent a link in this strategy-to-stakeholder-to-outcomes chain. In the transmission siting example, one record would capture the strategic priority: expand transmission in the target region. Another record captures the specific objective: siting at the target location. Another record describes an action that must be taken by a stakeholder outside the organization: approve the siting request. Yet another record represents the stakeholder who must take that action: the local city council. There are records for the communication and engagement activities meant to impact the stakeholder's posture: press releases, local radio interviews, staff meetings, briefings, town halls. Finally, there are records for what should be communicated during various activities: talking points for individual stakeholders as well as for specific issues.
Without AI, the platform becomes too cumbersome and too slow to keep up with a single complex strategy-to-stakeholder-to-outcomes chain; now multiply that by the number of chains that an organization may be focused on at any given time. Without the analysis and synthesis capabilities, reporting would be flat and miss key issues of concern before they become existential problems.
Scale preserves the human edge
Scale means the organization can manage more strategic priorities, stakeholder dependencies, evidence and coordinated cross-functional activity without a proportional increase in surveillance, reconciliation and reporting work. It does not mean producing more content. It does not mean eliminating practitioners.
AI prepares, connects, recommends and synthesizes. People validate evidence, interpret conditions, decide posture, exercise judgment, take responsibility and engage stakeholders. The platform can connect a meeting note to a commitment; it cannot own the relationship in which the commitment was made. It can surface an inconsistency; it cannot decide which operational tradeoff is warranted.
Preserving human responsibility is not a concession to technical limits. It is a condition of dependable operation. Evidence becomes useful only when someone with domain knowledge judges its relevance; a posture assessment becomes actionable only when someone accepts accountability for it.
We did not add AI to SCI because the market expects an AI story. We added it because the operating model would not scale without it. AI carries part of the recordkeeping and analytical burden; people retain the work that demands judgment, accountability and relationships. That is how AI helps strategic communication scale.
Notes
Russell Dubner, "Corporate Comms Is Playing Catch Up on AI. A Few Leaders Are Showing the Way," Boston Consulting Group, March 23, 2026. ↩︎
Mark Hutcheon, Alice Rutherford, Mariam Baffoe, and Ellie Laniyan, The Road to 2030: A Study of Corporate Affairs Functions in an Unpredictable World (Deloitte, 2026), 5, 7, 43–44. ↩︎
Oxford University Centre for Corporate Reputation and GlobeScan, Oxford–GlobeScan Global Corporate Affairs Survey: 2026 Survey Report (July 2026), 21–22. ↩︎
Ritoban Mukherjee, "What Is OpenClaw? Agentic AI That Can Automate Any Task," TechRadar, March 30, 2026; Darren Allan, "Mac mini Shortages Are Starting to Happen—and the OpenClaw AI Boom Is a Key Reason," TechRadar, February 17, 2026. ↩︎
OpenAI, "OpenAI and Hugging Face Partner to Address Security Incident during Model Evaluation," July 21, 2026; Stephen Witt, "Inside OpenAI's Hack of Hugging Face," The New Yorker, July 30, 2026. ↩︎