Skip to content
Partner service

Abacum Intelligence on a planning model built from drivers the business owns.

Constancia, an EPM consultancy and official Abacum (FP&A software) partner, reads your drivers, your actuals feed and your hierarchies before the layer is switched on. A person in finance signs what it produces.

Abacum at a glance

Covers
FP&A: budgets, forecasts, scenarios and reporting.
Data in
More than 700 integrations, with data arriving in near real time.
Our record
Certified architects and consultants based in the UK.
  • SOC 2 Type II
  • ISO 27001
  • GDPR compliant

Abacum Intelligence is the AI layer of Abacum (FP&A software). It forecasts, classifies, flags anomalies and branches scenarios from the planning model a finance team already runs. Constancia, an EPM consultancy, switches it on for finance teams whose board has asked for AI by a date.

Reads
The drivers, the actuals feed and the hierarchies already in the Abacum model, and nothing outside it.
Produces
A forecast, a set of anomaly flags, a cleaned load and a scenario branch, each traceable to the driver behind it.
Then
A named analyst signs each output before the board reads it, and the price of the checks is fixed after the survey.

Abacum Intelligence reads whatever the planning model holds, so we check the model before the layer goes on.

Abacum Intelligence covers 7 jobs, from cleaning the data and spotting anomalies to building scenarios and running the forecast. Each one reads the drivers, the actuals and the hierarchies already in the model.

The layer takes the manual work out of keeping a model trustworthy and current. It works hardest where the drivers have owners and the actuals tie to the ledger, which is what the survey settles before the layer goes on.

Where the headcount plan in the model and HR's payroll file disagree on 12 roles, the layer's forecast carries the 12 with full confidence. The FP&A manager then explains a salary line nobody can trace.

Abacum Intelligence hands 5 outputs to the FP&A team, and an analyst signs each before the board sees it.

Each output records the drivers it read, so the analyst who signs it can answer for it in the board meeting.

  1. 01

    The forecast

    The layer rolls the forecast forward from the drivers in the model, and the analyst reads it against last month's actuals before it goes anywhere.

  2. 02

    The anomaly flags

    A cost line that moves outside its usual range is flagged the day the actuals load, for a person to judge.

  3. 03

    The cleaned load

    Actuals arriving under a department the plan does not know are classified to the plan's hierarchy, with the mapping kept for the auditor.

  4. 04

    The scenario branch

    The layer builds a scenario branch from the driver changes the analyst names, and the base version stays untouched.

  5. 05

    The backsolve

    The layer works back from a board target to the driver values that reach it, and the analyst judges whether sales will own them.

The forecast is the output most teams switch on first, and the anomaly flags are the one they keep.

We make 4 checks on the planning model before Abacum Intelligence is switched on.

The 4 checks come out of the survey, and the consultants who made them switch the layer on.

  1. 01

    Every driver has an owner outside finance

    Headcount belongs to HR, pipeline to sales and volume to operations, and each owner has agreed to stand behind the figure.

  2. 02

    The actuals load from the ledger every month

    The connector runs on a schedule and the loaded balances tie to the trial balance, so the layer learns from consistent figures.

  3. 03

    The hierarchies match the consolidation

    Entities, accounts and departments in the plan are the ones the group result uses, so a forecast by entity can be checked against the close.

  4. 04

    A named person signs

    1 analyst in finance signs each output that reaches a board paper, and the model records the drivers it used.

Where a check fails, the fix goes into the fixed price before the layer goes live, and the survey names who owns it.

Abacum Intelligence stops at the judgement, and 3 things stay with the finance team.

The consolidation stays in the close, because Abacum plans and the consolidation stays where it already runs. The group result the forecast is measured against comes from there, and the layer reads it as a fact.

The choice between 2 gross margin definitions stays with the finance director. The model then carries the 1 chosen definition under that name, and the layer forecasts that 1.

The signature stays with an analyst, whose name sits on the board paper. The model keeps the drivers behind each figure, so the analyst can answer for it.

Abacum Intelligence reads a model that the implementation built and the connectors keep fed.

The layer is the last piece switched on, because it reads everything the other 2 pieces of work put in place.

A group on OneStream (EPM software) reads the SensibleAI page instead, because the layer there reads the consolidated model.

See more on our

Subscribe on YouTube
  • Latest episode

    How an EPM implementation works, and why the software comes last

Every Sunday, 6am

The Sunday Close.

Our weekly newsletter on AI in finance.
Yours to enjoy with your Sunday morning coffee.

  • AI in finance observations from Alex (CEO) and Brad (COO)
  • Newest video from our YouTube channel
  • AI in finance trends we have spotted each week

Tell us the date you are working to.

The people who would do the work read every message and reply inside 2 working days.