# Portal26 Adds Natural Language Capability to Platform for Tracking AI Impact > Source: > Published: 2026-09-09 13:05:59+00:00 TL;DR — Key Takeaways - Portal26 added **Recon** , a natural language query capability designed to surface AI governance, performance and cost issues using plain English. - Recon uses an **MCP server** and can help organizations track AI usage, tool calls, systems accessed and compliance concerns. - The platform can generate scorecards tied to KPIs and workflows while also applying detection rules as regulations change. Portal26 today extended its platform for tracking consumption of artificial intelligence (AI) resources to include a natural language query capability that makes it simpler to interrogate data captured by its platform. The Recon tool developed by Portal26 makes it simpler to, for example, discover governance, performance and cost issues using plain English rather than having to master a query language, says Neil Cohen, vice president of marketing for Portal26. Enabled via a Model Context Protocol (MCP) server, that capability is critical because most organizations lack any meaningful visibility into how much AI is actually being used for what purpose within their organization, he adds. “Everyone is flying blind,” says Cohen. Recon addresses that issue by surfacing insights into everything from how AI is impacting specific key performance indicators (KPIs) to compliance with specific mandates, notes Cohen. Additionally, Portal26 has templates for generating scorecards that track the impact AI is having on specific workflows and processes, he adds. Finally, Recon also closes detection gaps and continuously tightens security posture, including any time a new regulation is enacted using detection rules that can be deployed the same day. At its core, the Portal26 platform is designed to automatically discover and analyze usage of AI, including AI models, volume of tool calls, and systems being accessed, to enable organizations to better assess risk levels and other issues that might adversely impact performance or total costs. That latter issue [has become much more significant as organizations move to deploy AI agents that tend to sharply increase consumption of the number of tokens that providers of AI models use to charge organizations to invoke their services](https://techstrong.ai/features/portal26-adds-platform-to-govern-ai-agents/). At this juncture it’s not so much when governance will be applied to AI as much as it is how soon. Just about every organization that has adopted AI is now trying to get a better handle on everything from determining returns on investment (ROI) to whether AI agents might be accessing sensitive data without express permission. The only way to gain those insights is to deploy a platform capable of capturing the signals that AI tools, platforms and applications generate, notes Cohen. Ultimately, AI costs will eventually force a governance reckoning. While organizations were initially keen to drive as much AI adoption as possible for fear of being left behind by rivals, organizations are now focusing on identifying specific areas where AI provides the most business value, says Cohen. The issue is that rather than providing a competitive advantage, many capabilities provided by AI are now becoming table stakes that organizations will be required to have simply to remain competitive. One way or another, AI will soon be pervasively used across the enterprise with or without official permission. The challenge and the opportunity now is determining where and to what degree to apply AI in a way that doesn’t wind up breaking the budget for little to no real gain.