Agents / The Planner
Every scenario comes back as a full set of financials
Reforecast from live actuals, and model any operational driver you want to change. Either way the Planner returns a full plan: P&L, balance sheet and cash flow, regenerated across every entity, in minutes.
Runs on the same connected model as your reporting, so plan and actual never drift apart.
Tuesday 09:14 · Planner
Scenario · delay sales hiring one quarter
4 entities · FY26
EBITDA
19.7%
▲ 1.3 pts
Runway
16.3 mo
▲ 2.1 months
Revenue
£2.51M
▼ £410k
Statements regenerated
Profit and loss · 4 entities · rebuilt
Balance sheet · 4 entities · rebuilt
Cash flow · 4 entities · rebuilt
Consolidated group · eliminations applied
New entity go live · Q3 → Q4
Your plan is always slightly out of date, and always slightly out of reach.
Reforecasting takes a fortnight, so it happens quarterly at best. Modelling a decision takes days, so it happens after the decision. Both problems have the same cause: rebuilding a plan by hand is slower than the business moves.
Forecasts go stale on delivery
Two weeks of consolidating, chasing and rebuilding, and by the time it is signed off the actuals have already moved. The plan is a snapshot of a month that has ended.
Every scenario risks the real plan
Modelling in the live workbook means someone forgets to undo it. Modelling in a copy means the copy is stale within a week, and now nobody is sure which version the board actually saw.
You model one number, not the plan
There is time to work out the payroll saving. There is never time to follow it through to cash, covenant headroom and the balance sheet, so leadership decides on a partial answer.
Six jobs, from the live forecast to the decision it informs
Rolling forecast, reforecast and scenarios, all on the same connected model your reporting comes from. Not a slider on a dashboard.
Every account projected forward, continuously
The Planner carries each account forward from your actuals as they land, so the forecast is never older than your last refresh. No collection cycle, no consolidation week, no version to distribute. It also watches for drift, and tells you which accounts are pulling away from plan while there is still a quarter left to respond.
4 accounts drifting. Marketing spend is tracking 11% above plan and has been for three months.
Rolling forecast · 12 months
Updated with every actuals refresh
Flagged before quarter end
A full reforecast in minutes, not a fortnight
When the picture changes materially, rerun the whole forecast rather than patching the old one. Actuals to date, revised assumptions, and the remaining months rebuilt across every entity and consolidated to group. Reforecast because the business moved, not because the calendar said it was time.
Reforecast complete. 9 months actual, 3 months projected, four entities consolidated. 6 minutes.
Reforecast · FY26
Prior forecast retained as a version
Assumptions carried forward or restated
Change one operational driver, see the whole financial impact
Headcount, price, volume, lead time, churn, capacity, supplier terms or the date a new entity goes live. Change the operational number and the Planner works out every financial line it touches, including the ones you would not have thought to check. You describe the business decision. It handles the model mechanics.
Delay sales hiring one quarter. Six drivers move, three statements rebuild, four entities consolidate.
Drivers identified · 6
Drivers touched · 6
Dependencies resolved automatically
Three statements, every entity, consolidated
Not one metric recalculated. The Planner regenerates the profit and loss, the balance sheet and the cash flow for each entity, applies your elimination and FX rules, and consolidates the group. A scenario that only moves EBITDA is not a plan. It is an estimate.
P&L, balance sheet and cash flow regenerated across all four entities in under two minutes.
Regenerated · FY26
Intercompany eliminations applied
FX translation · close rate
Time to regenerate · under 2 minutes
Every version side by side, base plan untouched
Scenarios are versions, not edits. Run as many as the question needs, compare them on the measures that matter, and keep the ones worth keeping. The plan of record does not move until someone decides it should. Nothing you model can damage the plan you are already committed to.
Delay wins on margin. Expansion wins on growth. Both stay inside covenant.
Scenario comparison · 3 versions
Plan of record unchanged
Tells you what the decision actually turns on
Every scenario comes back with the numbers and with the thing you did not ask about. The dependency that breaks, the covenant that tightens, the date that stops working. The constraint is usually the answer. The useful part of a scenario is the part you had not thought to model.
The binding constraint is the new entity go live date, not the headcount. Delaying past Q3 breaks the launch timeline.
Constraints and dependencies
Traced to · headcount plan, go live milestones
Sources · HR system, project plan, ERP
Drill to source rows →
The forecast keeps itself current. You step in to test a decision.
The forecast rolls forward
Every account projects forward from actuals as they land. Drifting accounts are flagged without anyone asking, so the baseline is always current.
Ask, or reforecast
Change an operational driver to test a decision, or rerun the full forecast when the picture has moved. Both start with a sentence, not a template.
Get the full plan back
Three statements per entity, consolidated to group, with your eliminations and FX applied. Saved as a version alongside your plan of record.
Commit or discard
Promote a scenario to the plan when leadership decides, or leave it as a version you can point back to. Either way, there is a record of what changed and why.
A scenario should be safe to run
Most teams model less than they should because modelling feels risky. Versions get overwritten, assumptions go undocumented, and nobody is quite sure which file the board saw. The Planner is built so that running one more scenario costs you nothing.
Read how versioning works- Your base plan is never touched. Scenarios are separate versions. The plan of record only changes when a person promotes one, and you can always see who did.
- Every assumption is visible. Ramp curves, lead times, phasing and allocation basis are shown before the run and stored with the version, so a scenario from March still explains itself in July.
- Numbers trace back to source. Drill from any line in the regenerated statements to the driver, the rule applied, and the transactions underneath it.
- Enterprise security by default. Encrypted in transit and at rest, single tenant isolation, role based access down to report level, full audit logging, UK and KSA data residency, and aligned with GDPR and Saudi PDPL.
The same scenario answers a different question depending on your seat
Answer in the room, not in the follow up
When the board asks what happens if, you have the full picture rather than a promise to come back to them. The conversation stays a decision instead of becoming an action item.
Model freely without risking the plan of record
Every scenario is a version with its assumptions attached. No overwritten files, no orphaned copies, and a clear audit trail when someone asks how the number was arrived at.
Run five scenarios in the time one used to take
The rebuild is handled, so the work becomes choosing which questions are worth asking and pressure testing the assumptions behind them.
The Planner builds and maintains your forecast. Two more build the model and watch it.
Run the Planner on its own and your forecast keeps itself current, with any scenario a sentence away. Add the other two and the model underneath it builds itself, and tells you when something needs reforecasting.
The Modeler
Turns the workbook your team already trusts into a live connected model. It builds the foundation your forecast runs on.
Meet The ModelerThe Analyst
Watches the business for anomalies, gaps and cost you are carrying for no reason. Every insight arrives with a recommended action, and often with a reason to reforecast.
Meet The AnalystWhat finance teams ask us
How often does the rolling forecast update?
With every actuals refresh. Each account is projected forward from what has actually happened rather than from the last forecast cycle, so there is no point at which the numbers are waiting for someone to rebuild them.
How does it decide an account is drifting?
It compares the projection to plan and to the account's own history, and flags sustained directional movement rather than a single noisy month. You set the thresholds that matter to you, by account or by category.
How is this different from copying the file and changing a number?
A copy gives you one changed number and a file that is stale by next week. The Planner regenerates all three statements for every entity, applies your eliminations and FX, keeps the assumptions attached to the version, and stays connected to live actuals. The copy does none of that.
Which operational drivers can we model?
Anything the model is connected to: headcount and hiring phasing, price, volume, churn, capacity, cost per unit, lead times, supplier terms, CAPEX timing and entity go live dates. The point is that you change the operational number and the financial consequences follow automatically.
Can we set the assumptions ourselves?
Yes, and you should. Ramp curves, hiring lead times, phasing and allocation basis are all editable before the run and stored with the version. Where the Planner has assumed something you did not specify, it tells you what it assumed rather than burying it.
How many scenarios can we keep?
As many as the question needs. They are versions rather than files, so there is no proliferation problem to manage, and comparing five is the same effort as comparing two.
What happens when we decide to go with one?
Someone promotes it to the plan of record. The previous plan is retained as a version, the change is logged with who made it and when, and reporting picks up the new plan from the next refresh.
Does it plan for things other than revenue and cost?
Yes. Workforce, CAPEX, cash flow and balance sheet planning run on the same model, which is what allows a headcount decision to show up as a cash position rather than only as a payroll number.
Do we need the other agents for this to work?
No. The Planner needs a connected model, which the Modeler builds, but once that exists it runs on its own. Teams that add the Analyst tend to run more scenarios, because the findings surface questions worth modelling.
Bring your stalest forecast, or the decision you are stuck on
Twenty minutes, your model, and either a full reforecast or the scenario your leadership team has been arguing about. You will leave with three statements and an answer.