Lifesight today launched an MCP connector that lets CMOs and finance leaders query their live marketing measurement models inside Claude and ChatGPT - no analyst required, no API tokens needed, and no code to write.

Lifesight, the company behind what it describes as an agentic unified marketing measurement platform, today launched Lifesight MCP, a connector built on the Model Context Protocol (MCP) that gives Claude and ChatGPT direct, secured access to a company's live measurement model. The announcement was made on June 2, 2026, at 09:00 ET. According to Lifesight, customers collectively manage over $4 billion in marketing spend and operate across global markets.

The launch extends a pattern that has been building across the advertising technology industry since late 2024. MCP, an open standard introduced by Anthropic in November 2024, standardizes how AI assistants connect to external data and tools. From Google's open-source Ads API MCP server in October 2025, to Meta's AI Connectors for campaign management in April 2026, the protocol has spread through ad tech at a pace that few anticipated. Lifesight MCP now brings that same interface pattern to the measurement layer - the part of the stack where budget decisions are ultimately grounded.

What Lifesight MCP does

According to Lifesight, the connector allows marketing and finance teams to query their live measurement model directly inside the AI assistants they already use. Rather than waiting days or weeks for analyst-generated reports, CMOs, Heads of Growth, and finance leaders can pose questions in natural language and receive answers in seconds, framed around the business decision they are trying to make.

The technical architecture is deliberately constrained at launch. Lifesight MCP is read-only. Connector permissions map to the user's existing workspace access within Lifesight, meaning a user can only query data they are already authorized to see. No API tokens are required, no configuration files need to be edited, and set-up is completed through OAuth using an existing Lifesight login. According to Lifesight, the connector is available immediately to all Lifesight customers in Claude (Desktop) and ChatGPT (with Custom Connectors).

The platform supporting the connector serves more than 500 global brands. Every request made through the MCP connector is recorded in the customer's Lifesight audit trail, and the connector inherits the Lifesight platform's existing compliance posture - covering GDPR, HIPAA, SOC 2, and ISO 27001.

A Skills Library for marketing and finance workflows

Alongside the core connector, Lifesight today launched a Skills Library designed around common marketing and finance workflows. According to Lifesight, the launch capabilities include:

  • Board Briefing - creates board-ready summaries
  • P&L Translator - reframes marketing performance in the language of profit and loss
  • Scenario Planner - models alternative investment paths
  • Q4 Reallocation Stress Test - tests the robustness of spending plans under pressure
  • Channel Deep-Dive - provides granular performance breakdowns by channel
  • Anomaly Triage - surfaces and contextualizes week-over-week performance changes

The Skills Library approach is notable because it targets decision-makers who have historically been one or two steps removed from the raw measurement data. Board briefings, P&L translation, and scenario planning are not analyst workflows - they are executive and finance workflows. Making them accessible through a conversational interface inside Claude or ChatGPT positions the tool as a bridge between technical measurement infrastructure and the people who act on its outputs.

According to Lifesight's roadmap, future plans include assistant-guided recommendations and workflow automation requiring explicit human approval - a formulation that keeps a human in the decision loop while expanding what the system can initiate.

The measurement bottleneck problem

"For a long time now, marketing measurement insights lived in notebooks and dashboards that only some specialists could read correctly," said Tobin Thomas, CEO and Co-founder of Lifesight. "Lifesight MCP changes who gets to ask the question. When a CMO or a CFO can query the model directly - inside the assistant they already work in - the distance between insight and decision collapses from weeks to minutes. That decision speed, compounded over a year, is a real competitive advantage."

The problem Thomas describes is structural. Historically, accessing measurement models required routing requests through analysts or specialized teams. That bottleneck limited not only the decision-making process but also the number of questions that could realistically be asked. According to Lifesight co-founder and Chief Product Officer Rajeev Nair, "The hardest part was never the model - it was getting the right answers and decisions out of it. Every question used to be routed through a small team, which meant most questions never got asked at all. Lifesight MCP opens the model up to everyone who needs it. The people closest to the decision can finally ask their own questions and act on the answers - and the experts get their time back for the work that actually needs them."

That framing puts Lifesight MCP in the same conceptual category as a growing set of tools that have been trying to move marketing analytics from a specialist function to a general one. AppsFlyer launched its own MCP tool for mobile marketing measurement in July 2025, focusing on natural language queries against attribution data. Adverity debuted an AI-powered intelligence layer in September 2025 built on MCP technology, with an initial implementation automating data preparation for marketing mix modeling in Google Meridian. Newton Research integrated specialized marketing measurement agents with Snowflake Cortex AI in November 2025, enabling brands to run media mix modeling and incrementality analysis within secure data environments.

What Lifesight's platform measures

Lifesight's underlying platform unifies three distinct measurement methodologies. The first is marketing mix modeling (MMM), a statistical technique that uses regression analysis on historical time-series data to estimate the contribution of different marketing inputs to business outcomes. The second is geo-based incrementality testing, which measures the causal lift from marketing activity by comparing exposed geographies to control geographies. The third is incrementality-adjusted attribution, which attempts to combine the granularity of touchpoint attribution with the causal rigour of experimental methods.

According to Lifesight, this combination gives marketing and finance teams one trusted, causal view of what drives business growth, with the ability to forecast outcomes and optimize budgets. According to Lifesight, brands on the platform have achieved at least a 20% improvement in return on ad spend (ROAS).

The convergence of all three methods in a single platform matters because each answers a different question. MMM is suited to long-range portfolio planning and budget guardrails. Incrementality testing measures the true causal impact of specific campaigns or channels. Attribution provides granular, touchpoint-level reporting for operational optimization. The IAB and Instacart paper published in April 2026 argued that legacy measurement frameworks - those relying primarily on last-click or view-through attribution - systematically undervalue media channels that influence purchase decisions earlier in the funnel. Lifesight's unified approach is designed to address exactly that gap.

MCP's spread across ad tech

The broader context matters. MCP was introduced by Anthropic in November 2024 as an open standard for connecting AI assistants to external data sources and tools. Adoption across the advertising and marketing technology sector has been rapid. Google explored an MCP server for its Ads API as early as July 2025, then released an open-source implementation in October 2025. Google Analytics released an experimental MCP server for natural language data queries in July 2025. Amazon Ads launched its MCP Server in closed beta in November 2025, supporting connections from Claude, ChatGPT, Amazon Q, and Amazon Bedrock.

AdRoll and PubMatic demonstrated in 2026 how MCP can connect AI agents across two separate platforms to diagnose programmatic deal delivery failures in real time. Beehiiv launched an MCP integration in March 2026 allowing publishers to manage newsletters through Claude, ChatGPT, or Gemini. LiveRamp and Akkio partnered in April 2026 to bring conversational AI to marketing measurement, adding a natural language interface to LiveRamp's identity and measurement stack.

The pattern across all of these integrations is consistent. Enterprise systems that previously required specialized interfaces, dedicated analysts, or custom dashboards are being connected to general-purpose AI assistants through a standardized protocol. The AI assistant becomes the primary interface; the enterprise system becomes a data and tool provider behind it.

Lifesight's MCP launch fits that pattern precisely. The measurement model - previously accessible only through the Lifesight platform itself - is now queryable through Claude Desktop and ChatGPT Custom Connectors. The Skills Library bundles that access into pre-configured workflows for the executive and finance use cases most likely to generate direct business decisions.

The read-only constraint and what comes next

The decision to launch as read-only is worth examining. In the context of a live marketing measurement model, write access would imply the ability for an AI assistant to initiate budget reallocations, adjust model parameters, or trigger reporting workflows. That capability introduces accountability questions that most enterprise software vendors are currently approaching with caution.

According to Lifesight, future roadmap plans include assistant-guided recommendations and workflow automation requiring explicit human approval. The phrasing is careful: the AI proposes, a human approves. That architecture mirrors how other agentic advertising tools have been introduced. LiveRamp's agentic orchestration, launched in October 2025, gave AI agents access to its platform but maintained governance through permissioned access. The IAB Tech Lab's CEO, writing in late 2025, warned the industry against rushing automation without adequate governance frameworks in place.

The compliance posture of Lifesight MCP - GDPR, HIPAA, SOC 2, ISO 27001 - and the audit trail requirement for every request suggest that Lifesight has designed the connector with enterprise procurement processes in mind. Companies in regulated industries, including financial services and healthcare, routinely require demonstrable audit trails for data access. Building that into the MCP connector from launch, rather than as a later addition, removes a common barrier to enterprise adoption.

Prior Lifesight announcements in 2026

Lifesight MCP is not the company's first major announcement this year. In March 2026, Lifesight launched Mia, a marketing intelligence agent designed to help ecommerce and retail marketing teams analyze performance data, generate strategic recommendations, and act on insights across channels. Mia was built on Lifesight's platform and allowed teams to interact directly with their performance data through natural language. The MCP connector extends that conversational approach into the AI assistants that marketing and finance professionals are already using as their primary work interfaces, rather than requiring them to operate inside a dedicated Lifesight application.

According to Lifesight, the distinction matters. Mia operates within Lifesight's own environment. Lifesight MCP operates within Claude and ChatGPT - tools that many enterprise workers already have open on their desktops. The friction reduction of meeting people in the tool they are already using, rather than asking them to open a separate platform, is the structural bet behind the MCP approach across the industry.

Timeline

Summary

Who: Lifesight, an agentic unified marketing measurement platform serving more than 500 global brands, announced the launch. Tobin Thomas, CEO and Co-founder, and Rajeev Nair, Chief Product Officer and Co-founder, are the named executives behind the announcement.

What: Lifesight MCP - a read-only connector built on the Model Context Protocol - that gives Claude (Desktop) and ChatGPT (with Custom Connectors) direct, secured access to a company's live marketing measurement model, including causal marketing mix modeling, geo-based incrementality testing, and incrementality-adjusted attribution. The connector includes a Skills Library covering six pre-built workflows: Board Briefing, P&L Translator, Scenario Planner, Q4 Reallocation Stress Test, Channel Deep-Dive, and Anomaly Triage. Set-up requires no code and uses OAuth via an existing Lifesight login.

When: June 2, 2026, at 09:00 ET. Available immediately to all Lifesight customers.

Where: Available inside Claude (Desktop) and ChatGPT (with Custom Connectors). Lifesight is headquartered in New York, with a global brand base operating across international markets.

Why: Lifesight MCP addresses a structural bottleneck in marketing measurement: the gap between when insights are generated and when decisions are made. According to Lifesight, most measurement workflows require routing questions through analysts, which limits both the speed and volume of decisions. By embedding the measurement model directly into the AI assistants that marketing and finance professionals already use, Lifesight aims to reduce that cycle from weeks to minutes. The launch reflects a broader shift across enterprise software, where standardized protocols like MCP are enabling AI assistants to function as operational interfaces to backend business systems - a pattern that PPC Land has tracked across advertising and marketing technology throughout 2025 and into 2026.