BG-Demo is a fictional company. Every organisation, person and figure in the demo is invented, and everything shown is B-Glass's real output on that data.
The demo is a working B-Glass for BG-Demo, a London data business founded in October 2022 whose first four years, up to Fri 9 Oct 2026, were simulated.
Every number, date, name and relationship comes from a deterministic simulation: deals won and lost, renewals, invoices and payments, the costs in the ledger and the budget, platform usage, support tickets, staff joining and leaving. AI writers then turn those facts into what a real company holds (call transcripts, email threads, tickets, contracts, order forms and notes) using only the facts they are given, and checks catch any text that contradicts the record.
B-Glass then runs on BG-Demo exactly as it would on your business; the setting (cities, industries, regulation) is real.
Customer briefingBefore a meeting, one page on the account: its status and money, usage, open issues, the last calls and what was agreed, who matters on their side and who has joined or left, and the news about them.
Calls to makeIt spots the calls worth making, for a problem or an opportunity, and prepares each one with a briefing built from all the data and the actions agreed on earlier calls.
ProposalsIt sees what is working across your current customers, finds the prospects that look like them, and drafts the proposal.
Board reportThe quarter's numbers with the commentary written, as a deck: ARR, retention, pipeline, cash and the risks, each figure naming its source.
Churn and retentionWhy customers left, who looks like them now, and what to do about it before their renewal comes round.
Product usageWho uses what they pay for, what is growing and where to act, by customer, product and month.
What the demo holds
225 organisations
862 people at organisations
58 staff
1,995 calls
3,013 email threads
1,159 support tickets
1,218 documents
285 invoices
119 news items
Data as of Fri 9 Oct 2026
What the demo holds, by team
The boardThe Commercial Dashboard: ARR, growth, retention, margins, Rule of 40 and headcount, each against its target, with the AI's commentary.
FinanceThe P&L by month with costs by category, budget against actual, debtors and cash, and invoiced ARR reconciled against the CRM.
SalesWon deals and pipeline by owner against target, renewals coming up, and every call and email.
Customer successEvery customer on one page: the AI's assessment, usage against what they pay for, tickets, documents and the next step.
PeopleStaff by department, each with a page on their role, when they joined or left and the accounts they hold; who reports to whom; time off and whereabouts.
Invoices, deals, payments and the ledger are not read one by one: they feed the facts behind every figure, and the AI's summaries and commentary are written from those facts.
Layer by layer
1. It reads every object on its ownEach call, email thread, support ticket, document and news item is read by the AI, and what it found sits on that object's own page, so you can drill in from anywhere.
2. It rolls up by organisation and personEach organisation gets an assessment, a status, next steps and its story; each person at an organisation an insight; each member of staff a profile; each with the reasoning that fits it.
3. It writes the summaries at the topThe Summary's Executive overview and This week, and the commentary on the Commercial pages: the accounts, the money and the people together.
4. It maps every documentEvery contract, order form, deck and note is read and filed by organisation and kind on a document map; each document's page shows what was found in it and where (As found in this document).
5. You can trace it all backThe Evidence trail follows any sentence or figure down through those layers to the source rows it came from.
A short tour
A picture for each team: the week's calls and a customer for sales and customer success, the Dashboard for the board, the P&L for finance, People, and the evidence behind them all.
The Summary
Demo: BG-Demo, a fictional company
Every call of the week, read by the AI and laid out by day, with the week ahead under it.
A customer's page
Demo: BG-Demo, a fictional company
Brackwell Freight's page: its status, ARR and renewal in the title, and what the AI read from every call, email and ticket beside the profile.
The Commercial Dashboard
Demo: BG-Demo, a fictional company
Every kpi from one set of facts: ARR, retention, margins, efficiency and headcount, each against its target.
The Profit and Loss
Demo: BG-Demo, a fictional company
Revenue against the costs in the ledger, by category, with the AI's commentary on EBITDA, margins and what is off target.
People
Demo: BG-Demo, a fictional company
Every member of staff by department, with the active count and who has left; time off and who reports to whom sit beside it.
The evidence trail
Demo: BG-Demo, a fictional company
One AI sentence picked, Brackwell Freight at risk, with every record it was built from and every page that shows it.
How to use it
Find and Ask B
Find, in the header, searches everything in your browser.
Ask B answers in plain English with links to its sources. It runs on Claude on the demo machine, SILVER, so it works only while SILVER is on, and it asks once for the demo key: ask the person who gave you this link. When SILVER is off it offers the Claude Project instead.
The clay Claude button opens the B-Glass Project in Claude, which answers from the same data.
Every card ends with a made from link that opens the evidence trail for that card. The theme button at the top right switches dark and light; all data is as of Fri 9 Oct 2026.