Role
Financial analyst
What do you not know about the next quarter until the business tells you itself? The forecast rests on assumptions that arrive verbally, late, and in a different form from every department.
Intro call
Four questions we would start with
What do you not know about the forecast until the departments tell you?
If it is practically everything, then the forecast is the sum of other people's declarations - and you are accountable for how accurate it is without any influence over a single input.
How many people touch one version of the budget?
With a dozen or so departments and one spreadsheet sent by email, there are as many versions as there are people - and nobody knows which one is in force.
How long does preparing a new scenario take?
If it is two days, then the board's question "and what if sales drop by ten per cent" gets answered a week later. The value of the answer falls with every day.
What happens when an assumption turns out to be out of date mid-quarter?
Usually: the forecast stays as it was, because recalculating costs too much. And from that moment everyone knows the plan is out of date, but nobody changes it.
We take over building the model, you keep the judgement
Collecting assumptions, versioning the budget, recalculating scenarios and comparing actuals against plan can be taken off the desk. What to assume and how to read the risk stays with the analyst.
- The forecast is not produced automatically. - The numbers are calculated by the model, but the assumptions are entered by a person and that person is accountable for whether they make sense. An automatic forecast without judgement is an extrapolation of the past.
- Every assumption has an author and a date. - When figures disagree it is clear who declared what and when - not in order to hold anyone to account, but so that it is possible to go back to the source.
- A scenario recalculates without rebuilding the model. - That is the change after which "what if" stops being a nuisance and starts being a tool.
Scope
What can be taken off financial planning
Twenty example tasks. Each one has a human approval point. Below is a selection, and the scope we fit to your process.
Collecting assumptions from the business
- Collecting the departments' declarations in one form, with the author and the date
- Reminders about assumptions not yet delivered
- Pointing out assumptions that depart from that department's history (assessment: the analyst)
- A register of changes to the assumptions during the period
- Comparing what a department declared with what it delivered in earlier periods
The model and the scenarios
- Recalculating the forecast after an assumption changes
- Preparing variant scenarios (choice of assumptions: the analyst)
- Pointing out which assumptions affect the result most
- Checking the model is consistent - whether the parts add up to the whole
- Documentation of the model produced along with it
The budget and its versions
- Versioning the budget with what changed between versions
- Consolidating the departmental budgets into one whole
- Detecting items planned twice in different departments
- The budget against last year with the differences explained
- Checking the budget fits within the limits set by the board
Variances and material for the board
- Actuals against plan after the period closes
- Breaking a variance down into price, volume and mix
- Pointing out departments that systematically over or under-estimate
- Material for the board meeting in a fixed form (conclusions: the analyst)
- An archive of forecasts with how they turned out
A day's work
Today
- You collect assumptions by email and in conversations, each in a different form
- Consolidating the budget is a week of work in spreadsheets
- A scenario the board asks for means rebuilding the model
- When figures disagree you cannot remember who declared a given number
- One person understands the model, and that person's holiday stops the planning
- You do not know how last year's forecasts turned out
With AI assistance
- The assumptions arrive in one form, with the author and the date
- Consolidation happens by itself, you check it is consistent
- You recalculate the scenario during the conversation with the board
- The history of assumptions is recorded, so a discrepancy has a source
- The model has documentation, so it can be handed over
- The forecast archive shows which departments are systematically wrong, and in which direction
- Time to build a scenario
the answer to "what if" comes during the conversation
recalculating stops requiring a rebuild of the model
- Trust in the assumptions
it is known who declared what and how it turned out for them
assumptions have an author, a date and a history
- A model you can see into
planning stops depending on one person
the model has documentation produced along with it
- Speed of reacting to change
an out-of-date assumption is corrected rather than passed over
recalculating costs minutes, so there is no reason to put it off
We start with collecting the assumptions, not with building the model.
- Step 1
One form for the assumptions
We establish exactly what you need from each department and in what shape. A simple form instead of a spreadsheet sent by email changes more than it seems.
- Step 2
Consolidation and versioning
Assembling the budget from the inputs, and a version history. It removes the manual assembling and ends the arguments about which version is in force.
- Step 3
Scenarios
Recalculating after an assumption changes. Only once the inputs are in order - otherwise you quickly calculate things nobody trusts.
- Step 4
Variances and the forecast archive
Comparing actuals with plan and judging accuracy. It needs several periods of history, so the value shows after a year.
Common questions
Questions that come up at the first meeting
We have a model in a spreadsheet that works. Why touch it?
If it works, we do not touch it - and that is the honest answer. The question is rather how many people can open it without breaking it, and what happens when the author goes on holiday halfway through the budget cycle. A spreadsheet one person understands is a risk however well it calculates. We usually start by consolidating the inputs around that model rather than by replacing it.
Will this role not be cut once the budget assembles itself?
A budget is made of decisions, not of spreadsheets. We take over the consolidation and the versioning, the work that takes a week and that nobody considers analysis. What is left is harder: the conversation with budget owners, judging the assumptions and asking whether the plan is feasible at all. In today's cycle there is usually no time left for that.
Does this fall under the AI Act?
Financial planning is not a high-risk use. The point to watch is elsewhere: a forecast used to decide about hiring or redundancies touches people, and separate rules apply there - in particular, such decisions are not based on a calculation alone.
Still have a question that is not here? Book a call
Next step
How long does an answer to "what if" take in your company?
We will show you what planning looks like when assumptions have an author and a scenario is calculated during the conversation.