Custom AI Apps
Software we built to do work we were doing by hand
Three products, each built inside a Blash engagement to do work we were otherwise doing in a workbook: finance operations, budgeting and outbound. Each is now available to clients.
12-week
cash forecast across every company in a group, alongside 4-week and 6-month views
Multi-entity
consolidation, with intercompany trade stated on the face of the numbers before eliminations
Plain English
questions answered in Budgeting from figures already computed, never estimated
Read-only
demo for each product, on synthetic data
01
What we make
One job each. Every product has a read-only demo running on synthetic data.
02
Why an advisory firm makes software
We did not set out to build products. Each of these began as a recurring problem on an engagement, and the alternative each time was another workbook that only one person understood.
Because they were built to do the work rather than to be sold, they are shaped by how the work is actually done. Budgeting states the intercompany trade on the face of the dashboard, one contract line the group sells to itself, and says plainly that consolidated the group earned neither side of it. Cashflow reports DSO on submitted date as well as invoice date, because for a lot of businesses those two numbers are weeks apart and only one of them is the real problem. Budgeting carries a gross margin floor, set at group level with companies allowed to differ, and flags the departments sitting below it.
Each one is configured before it is used: the companies in the group, the fiscal calendar, the departments and directorates, the routing an approval follows, and the definitions the numbers are built from.
03
What they have in common
GTM and Budgeting each carry AI, and in each it does one specific job you can watch it do.
In GTM it sorts a classified list against an ideal customer profile. In the demo list of 11,000 that is 45 per cent dropped, 35 per cent qualified and 20 per cent left as uncertain rather than forced into a decision. For each prospect it keeps, it finds a current signal about that person, picks the single service to lead with, and drafts.
In Budgeting you can ask the workspace a question in plain English and get an answer assembled from figures the model has already computed, with the provenance of the recompute attached. In its own words: a question the workspace cannot answer gets told so rather than guessed at, it never estimates a figure, and it never returns an individual's name or pay.
The AI reads, qualifies, drafts and explains. It does not invent numbers. Cashflow and Budgeting both run automated checks over their own figures, and Cashflow says on the dashboard when a cash or working capital balance has not been updated that day. A system that will say it does not know is worth more than one that always has an answer.
04
Where to start
Each product has a short tour and a read-only demo built on a fictional company. Nothing in them is a client result; the figures are illustrative fixtures. Open the one closest to your problem, then talk to us about the fit. Some engagements use these products directly. Others start from them and go somewhere else.
Common questions
Do I have to be a Blash client to use these?
No. Each product stands on its own and can be used without any advisory work from us. They sit alongside an advisory engagement well, but neither requires the other.
Are the numbers in the demos real?
No. Every demo runs on synthetic data for a fictional company and is read-only. The figures are illustrative fixtures chosen to show how the product behaves. They are not client outcomes and should not be read as any.
What does the AI actually do?
In GTM it qualifies a contact list against an ideal customer profile, drops the people who do not fit, finds a current signal for each person it keeps, chooses the service to lead with and drafts the message. In Budgeting it answers questions in plain English from figures already computed in the model. In both it reads, classifies and explains. It does not decide, and it does not produce numbers of its own. Cashflow has no AI layer; its forecasts, ratios and checks are computed.
Will it make a figure up if it does not know?
No, and this is deliberate. Budgeting tells you when a question is outside what the workspace can answer instead of guessing, and by its own statement it never estimates a figure. GTM keeps an uncertain bucket rather than forcing every contact into qualified or dropped. Cashflow and Budgeting both run automated checks over their own numbers, and Cashflow shows the failed checks and warnings on the dashboard next to the figures rather than behind them.
Do Cashflow and Budgeting work for a single company?
Both are built around a group. The consolidation, the intercompany eliminations and the single cash position across entities are the parts hardest to do by hand, so that is where most of the value sits. For a single company we would want to look at your case before saying yes.
Can they be adapted to how we work?
Usually, yes. Chart of accounts, entity structure, approval routing, the departments and categories a budget is built from, the definition of a qualified prospect: these differ by client and are meant to be set. Start with the demo, tell us where it does not match, and we will tell you whether it is a fit.
Next step
See it on your own numbers
A short walkthrough with the people who built it.
