From portfolio reporting to source-connected answers

The business need

A hospitality portfolio needs to understand performance at several levels. Leadership needs a view across the portfolio, market teams need to understand local movement, and property teams need to know which pages, audiences and acquisition journeys require attention. A change in traffic means little without the right period, source and commercial context.

Digital Squad developed a shared measurement product with a conversational AI assistant. The objective was to make the evidence accessible to the people responsible for acting on it, while preserving the distinctions that make performance reporting trustworthy.

A measurement system built around the operating model

We organised the experience around portfolio, market and property scopes, connecting Google Analytics and Search Console evidence with explicit metric definitions and reporting periods. Teams could move between an overall view and the supporting detail without treating every source as interchangeable.

Search exposure, website behaviour and commercial outcomes each answer different questions. Keeping their definitions and availability visible gives an investigation a sound starting point. Missing data remains distinguishable from a genuine zero, and interpretation stays connected to the source period it describes.

Client feedback also shaped the product. The experience moved toward plain-language explanations, questions worth investigating and next actions in the relevant market or property context. The reporting workflow became a product-design problem as well as a data-engineering task.

Engineering the analytical assistant

We built a custom assistant that can retrieve evidence through native, server-executed tools. It can discover the available scope and datasets, retrieve canonical metrics, query supporting breakdowns and consult recorded opportunities. This allows a performance question to lead into a focused investigation using the same measurement foundation as the reporting experience.

The application controls access and execution. Authorised scopes are resolved by the system, tool inputs are validated, and read-only queries have bounded results and execution limits. The assistant interprets retrieved evidence while metric calculations and access decisions remain under application control.

That separation is central to the engineering. A fluent answer must still refer to the correct property, reporting period and data source. The model has room to explain the evidence, with the application setting the boundaries of what it can retrieve.

What the work delivered

The work brought source-connected reporting and conversational investigation into one authenticated client-review experience. Portfolio and property teams gained a common measurement foundation, with an assistant designed to help them explore performance without assembling a new analysis from disconnected exports for every question.

The delivered capability is specific: structured access to performance evidence, contextual investigation and a clearer route from a reporting question to the next action. The client-review release also created a basis for feedback-led development of the wider reporting product.

Where this approach applies

For organisations managing several markets or business units, analytical AI depends on the quality of the measurement infrastructure beneath it. Digital Squad connects the operating context, data model and application controls so teams can investigate performance with consistent definitions and accountable access.