Section one
A report your AI can actually read.
We've built a small server that exposes every part of Pints & Profits 2026 — the four chapters, the seven key findings, the curated charts, and the raw n=1,002 survey crosstab broken down by age and gender — using the Model Context Protocol, an open standard for connecting AI assistants to live data sources.
Most reports sit as a PDF on a server. An AI assistant asked about them either makes something up, summarises a marketing page, or refuses. This one is different. Ask Claude — or any MCP-compatible assistant — a question about the UK pub and bar sector and it can answer using our research, not whatever the open web happened to scrape. Quote a specific statistic. Compare age cohorts. Pull a chart. Summarise a chapter for a board paper. It's the report, but conversational.
1,002
UK adults surveyed by Sensu Insights, October 2025.
10
tools your AI can call — search, chapter, chart, crosstab.
5
age groups, broken out by gender, on every survey question.
Section two
Try asking, for instance.
Paste any of these into your AI assistant once the server is connected. The first answer usually triggers half a dozen follow-ups.
What does the survey say about how 18–24 year olds discover pubs and bars? Which channels should we be investing in for that cohort?
Example prompt
The server answers from our data, so follow-ups can be specific: pull the underlying chart, drill into the crosstab, compare against a different age group, or summarise the relevant chapter for a different audience.
Summarise the “digital fault line” section for a board paper. Two paragraphs, three bullet recommendations.
Example prompt
Compare on-trade visit motivations between 25–34s and 55–64s. Where do they diverge most sharply?
Example prompt
Pull the chart on information discovery sources by age, and tell me what it implies for our marketing efforts.
Example prompt
None of these require any prompt engineering or special syntax. Just connect the server once, then ask in plain English.
Section three
Connect in under a minute.
The server lives at report-mcp.wearebrew.com — vanilla MCP over Streamable HTTP. No API key, no sign-up.
If you use Claude
In Claude, go to Customize → Connectors, click the + icon, and add a custom connector with:
- Name: Pints and Profits 2026
- URL:
https://report-mcp.wearebrew.com/
That's it. The ten report tools will appear in your conversation immediately.
Claude Desktop, Cursor, and other MCP clients
Edit your client's MCP server config and add this block:
{
"mcpServers": {
"pints-and-profits-2026": {
"transport": "http",
"url": "https://report-mcp.wearebrew.com/"
}
}
}Restart the client. You should see “pints-and-profits-2026” in the connected servers list. The MCP Inspector and any other MCP-compatible client work the same way: point them at the URL above and you're done.
Section four
Need a hand connecting your AI?
Setting up MCP connectors, briefing your team on how to use them, or putting the report's data to work in your own marketing — we're happy to help.
Whether you're an operator who wants help wiring this up for your team, an industry body looking to put the data in front of members, or a journalist who needs a tailored cut of the survey, get in touch and we'll point you in the right direction.
