THE FOREIGN SERVICE JOURNAL | MAY-JUNE 2026 41 deliberately will result in more mature solutions for a fully AI-integrated customer relationship management (CRM) platform from Microsoft and Salesforce, among others. I base this observation on my experience leading a procurement and global implementation of Salesforce for the U.S. Department of Commerce—a process that took more than two years and required significant executive sponsorship, specific statutory language appropriating funds, and a team of seasoned field officials and tech contractors. How can AI be helpful in the near future? Imagine having a team of people at post who speak requisite languages with fluency, who have the bandwidth to read every article and consume all public media, and whose job it is to generate relevant reports, receive and transcribe debriefs, and flag patterns and people of interest to U.S. diplomats who cycle through posts every two years. If an AI agent can do this, it frees our diplomats from the time-honored but questionable practice of translating news articles and sending them in as reports with a snippet of commentary. It enables them to focus on what an AI agent can’t do—engage on a personal level to promote U.S. national interests and gather actionable, relevant human intelligence. AI can already be deployed to read all public media and tag it for structured retrieval and synthesis. And since it feeds off public data, it could be implemented quickly as a pilot at a few posts. AI could also be used to derive reports, metrics, and prompts for action. Within Microsoft, such tools as CoPilot are already being deployed for writing at some U.S. government agencies (e.g., U.S. Export-Import Bank). AI implementation would start with new standards for presence at post, established by the ambassador with support from the Office of Management and Budget (OMB) to insist that common goals drive collaboration between country team members on a common technology platform. Incorporating the Human Touch Of course AI cannot replace human intelligence gathering, the relationships that embassy officials develop. Yet such information needs to feed into the CRM platform, and this could be vastly facilitated by AI, via dialogue. Most CRM implementations fail due to lack of user adoption. Users don’t like filling in data fields and won’t do so unless forced to; and if CRMs are not populated and maintained, they are useless. But what if an embassy officer went to a meeting with a contact and afterward simply initiated a conversational dialogue with an AI agent? The AI agent could be trained to elicit points that allow structured tagging to feed the CRM. “Whom did you meet? Is this the right spelling of their name and title? Do you have contact information? What issues did you discuss? Are other countries getting involved in supporting unrest over this mining investment? What are the next steps? When do they need to be taken and by whom?” And so on. This sort of conversation-to-text tool is now being used in the medical field to transcribe and summarize meetings with patients. And the capability is being developed by leading CRM platforms. AI can also be used to stimulate proactivity. It can search CRM casework and contact profiles for gaps—a vice-minister who has no documented U.S. embassy contact, or a private sector executive who hasn’t been contacted in three months, for instance. In concert with public news consumption, AI can create a prioritized listing of contact outreach for any embassy official. And it might be able to offer suggestions of important questions (e.g., “I saw that Brazilian firms may be in country seeking to assume that mining contract, can you confirm?”) based on those public sources and reporting from other embassy colleagues. The same AI can generate any number of structured reports for officers to read or listen to even before they arrive at post (e.g., how investment is being treated in Panama, what labor unrest there is, who the key government contacts are and how to engage them). The beauty of AI is that it is dedicated to being responsive to natural language requests. If the data is accessible, then AI tools can already generate better summaries and points than most leave-behind briefs I inherited from my predecessors at post. Outgoing officers are typically scrambling to move, so briefing their replacements isn’t high on the list of priorities. AIgenerated reports will give newly arrived officers a head start on maintaining and deepening relationships of interest. Shifting the Culture of Diplomatic Work In conversation with reporting officers, I’ve often heard exasperation with a 1950s mentality of reporting (e.g., “Cables aren’t saying much more than what I can find myself online”). How can we facilitate a shift in culture to acknowledge that the highest value added by reporting officers at post is offline human intelligence gathering and getting outcomes? What carrots and sticks need to be put in place, what sort of grading system? Here the application of metrics might be tricky, but academia points the way. If AI is already capturing all public sources and feeding a CRM with the knowledge of an individual or ministry, then it can flag unoriginality in a reporting officer’s submission (e.g., “This report is just a rehash of articles in the
RkJQdWJsaXNoZXIy ODIyMDU=