Automated workflow · n8n + Claude

Budgeting & Forecasting Projections

Feed in a line item's trailing four quarters of actuals and any known context, and get back a Q1–Q4 base/upside/downside projection for the coming year, plus one consolidated annual forecast per forecast group — split quarterly, with the assumptions boiled down and a short commentary for leadership, not a spreadsheet dump.

forecast groups

line items

annual roll-ups

Deterministic math, then Claude's judgment — twice

This pipeline computes deterministically at two points, not one. First, per line item: a code node turns four quarters of history into a growth rate and a naive trend baseline. Second, after every line item has its own quarterly forecast: a code node sums those quarterly bands across the whole group into one deterministic annual total. Claude never does the arithmetic at either point — it decides which forecasting method actually fits each line item, and later writes the judgment call on top of the group's already-computed totals.

Trend extrapolation

The naive baseline is a good fit — steady historical growth, no known disruptions. The base case tracks it closely.

Seasonal adjustment

History has a repeating shape a straight growth-rate would badly misread — a compounding trend line through a holiday spike, for instance. The base case follows the seasonal pattern instead.

Assumption-driven

The notes describe a known future change — a price hike, a launch, a renewal — that should shift the projection at a specific point, regardless of what history says.

Computed deterministically (2 code nodes)

Per line item: average growth rate and a naive quarterly baseline. Per forecast group, after every line item is done: the quarterly and annual base/upside/downside totals, summed straight across all line items. Same inputs always produce the same numbers.

Judged by Claude (2 AI Agent calls)

Per line item: which method fits, confidence, the quarterly scenario band, rationale, and a short stakeholder message. Per forecast group: overall confidence, the 3–5 assumptions actually driving the number, and a short annual commentary — on top of totals it never has to compute itself.

How a forecast request moves through it

One straight path from intake to per-item analysis, then the array of line items splits, fans out to the archive, and recombines into a group total that gets one more round of judgment before responding.

01 · trigger

Forecast Webhook

POST /webhook/forecast-projection receives a forecast label and an array of line items, each with four quarters of historical actuals.

→

02 · normalize

Normalize Input

Cleans the payload; defaults the horizon to 4 quarters and a missing line-item array to empty.

→

03 · deterministic

Compute Trend Baseline

For every line item: the average quarterly growth rate and a naive baseline compounding it forward through Q4 — no LLM involved yet.

→

Generate Forecast

Claude picks the method, sets confidence, builds the Q1–Q4 base/upside/downside band, and drafts a short message — per line item, in one call.

→

04 · fan out

Split Out & Flatten

Turns the one forecasted array into separate items, one per line item.

for every item, in parallel

Save Forecast

Archived to budget_forecast_projections, one row per line item.

→

Collect Line Items

Re-combines every item in the group back into one array.

→

05 · deterministic

Compute Annual Rollup

Sums every line item's quarterly band straight across the group into one quarterly and annual total — again, no LLM.

→

Generate Consolidated Forecast

Claude sets overall confidence, distills the key assumptions across all line items, and writes one high-level annual commentary.

fan out once more

Save Annual Rollup

Archived to budget_forecast_annual_rollup, one row per forecast group.

→

Respond With Forecast

Sends every line item's forecast plus the consolidated annual rollup back in one response.

Three forecast groups, six line items, one roll-up each

The companies, departments, and dollar figures below are made-up test data, not a real business's books. Historicals are mocked as four quarters of trailing actuals (FY2026); projections cover the four quarters of the coming year (FY2027). The trend-baseline and roll-up math are deterministic (not AI); every method choice, scenario band, rationale, message, and the consolidated commentary are genuine, unscripted Claude output from running each group through the live workflow.

SaaS Revenue Streams

FY2027 Q1–Q4 forecast

Recurring Subscription Revenue

Revenue

Assumption-driven

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$420K	$445K	$468K	$495K

Projection — FY2027 · confidence: medium

	Q1	Q2	Q3	Q4
Upside	$549.5K	$610.4K	$677.6K	$752.7K
Base	$534.6K	$577.8K	$624K	$673.9K
Downside	$519.8K	$545.8K	$573.1K	$601.8K

Rationale

The historical trend is steady at ~5.6% QoQ, which the naive baseline faithfully captures. But a new enterprise tier launches in Q1 FY2027 — a known future step-change that makes assumption-driven the correct method. The base case models an acceleration to ~8% QoQ from Q1 onward; the downside effectively treats the tier as a non-factor, tracking the historical rate instead.

Automated message · Slack/email to stakeholder

Forecast → Revenue leadauto-drafted

Recurring Subscription Revenue is projected to accelerate meaningfully in FY2027, with our base case reaching approximately $674K by Q4 — driven by the new enterprise tier launching in Q1. Growth is expected to outpace the historical ~5.6% quarterly rate, though actual trajectory will depend heavily on enterprise adoption speed. Confidence is medium given the launch is confirmed but uptake is still to be determined.

Professional Services Revenue

Revenue

Trend extrapolation

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$82K	$84K	$81K	$85K

Projection — FY2027 · confidence: high

	Q1	Q2	Q3	Q4
Upside	$88K	$90K	$92K	$94K
Base	$86K	$87.1K	$88.1K	$89.2K
Downside	$83K	$84K	$85K	$86K

Rationale

Actuals are a near-flat, low-volatility pattern with a 1.2% average QoQ growth rate, no seasonal shape, and nothing flagged in the notes for FY2027 — trend extrapolation is the right call, and the naive baseline is carried forward largely as-is.

Automated message · Slack/email to stakeholder

Forecast → Revenue leadauto-drafted

Professional Services Revenue is expected to remain stable throughout FY2027, tracking a gentle upward drift from roughly $86K in Q1 to $89K by Q4. No structural changes are anticipated, and the historical data shows very low volatility — giving us high confidence in this projection. The key driver is simply the continuation of the existing steady run rate.

Annual roll-up: SaaS Revenue Streams

FY2027 annual forecast

Confidence: medium

Consolidated across both line items above — quarterly totals computed deterministically, confidence and commentary judged by Claude.

Quarterly base-case total — FY2027 (sum of both line items)

	Q1	Q2	Q3	Q4
Recurring Subscription Rev.	$534.6K	$577.8K	$624K	$673.9K
Professional Services Rev.	$86K	$87.1K	$88.1K	$89.2K
Total (base case)	$620.6K	$664.9K	$712.1K	$763.1K

FY2027 annual base: $2.76M

Upside: $2.95M

Downside: $2.58M

Key assumptions

The new enterprise subscription tier launches on schedule in Q1 FY2027 with no material delay.
Enterprise adoption drives Recurring Subscription Revenue growth of ~8% QoQ — a meaningful step-up above the historical ~5.6% trend but well below the bull-case ~11%.
Professional Services remains structurally unchanged with no new initiatives, keeping that stream flat-to-modestly-growing throughout FY2027.

The portfolio is on an upward trajectory from Q1 through Q4, but the shape and magnitude of that growth curve depend almost entirely on one event: the Q1 FY2027 enterprise tier launch. Because Recurring Subscription Revenue dominates total portfolio value, the ~$376K spread between the upside and base case — and the ~$182K gap to the downside — is driven exclusively by how quickly enterprise customers convert and ramp, making adoption velocity the single most important variable to watch this year. Professional Services is a stable, low-volatility contributor that requires no active management attention. Leadership should establish a formal review gate at the end of Q1 using actual enterprise conversion data to confirm whether the model is tracking, or whether a re-forecast is warranted before H2 commitments are locked.

Retail Sales Lines

FY2027 4-quarter forecast

Holiday Retail Sales

Retail

Seasonal adjustment

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$180K	$95K	$88K	$310K

Projection — FY2027 · confidence: medium

	Q1	Q2	Q3	Q4
Upside	$237.5K	$125.2K	$116K	$408.9K
Base	$215.9K	$113.9K	$105.5K	$371.7K
Downside	$194.3K	$102.5K	$94.9K	$334.5K

Rationale

A classic seasonal retail pattern: a Q4 holiday spike (~46% of the year) followed by a sharp reversion to off-season levels. The naive trend baseline, driven by a 19.87% average growth rate, catastrophically misreads this — it would project runaway growth off the Q4 spike. Seasonal indices from FY2026 are instead applied to a FY2027 annual total grown by that same 19.87% YoY rate, preserving the seasonal shape.

Automated message · Slack/email to stakeholder

Forecast → Retail leadauto-drafted

Holiday Retail Sales is projected to follow its strong seasonal pattern in FY2027, with Q4 again expected to account for roughly 46% of annual revenue (~$372K base case) while Q2 and Q3 remain the natural trough quarters. The primary driver is the continuation of the observed seasonal shape combined with ~20% year-over-year underlying growth. Confidence is medium, as the seasonal index is based on a single year of history, so we'd recommend revisiting once Q1 actuals are available.

Everyday Apparel Sales

Retail

Trend extrapolation

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$140K	$144K	$149K	$153K

Projection — FY2027 · confidence: high

	Q1	Q2	Q3	Q4
Upside	$160K	$166.4K	$173.1K	$180K
Base	$157.6K	$162.3K	$167.2K	$172.2K
Downside	$155.3K	$157.6K	$160K	$162.4K

Rationale

A remarkably steady, near-linear growth trajectory at ~3% per quarter with no seasonal distortion — exactly the pattern trend extrapolation is designed for. The naive baseline aligns closely with this and is adopted as the base case.

Automated message · Slack/email to stakeholder

Forecast → Retail leadauto-drafted

Everyday Apparel Sales is expected to continue its steady ~3% quarterly growth trend through FY2027, reaching approximately $172K by Q4. The key driver is the consistent organic demand pattern observed across all four quarters of FY2026, with no known disruptions planned. Confidence is high given the low historical volatility of this line item.

Annual roll-up: Retail Sales Lines

FY2027 annual forecast

Confidence: medium

Consolidated across both line items above — quarterly totals computed deterministically, confidence and commentary judged by Claude.

Quarterly base-case total — FY2027 (sum of both line items)

	Q1	Q2	Q3	Q4
Holiday Retail Sales	$215.9K	$113.9K	$105.5K	$371.7K
Everyday Apparel Sales	$157.6K	$162.3K	$167.2K	$172.2K
Total (base case)	$373.5K	$276.2K	$272.7K	$543.9K

FY2027 annual base: $1.47M

Upside: $1.57M

Downside: $1.36M

Key assumptions

Holiday Retail Sales sustains ~20% YoY growth in FY2027, the single largest lever on the annual total.
The FY2026 seasonal distribution (Q4 ~46%, Q1 ~27%, Q2/Q3 troughs) repeats reliably in FY2027, despite being calibrated from only one year of actuals.
Everyday Apparel continues its stable ~3% sequential quarterly growth with no structural demand disruption.

The portfolio is sharply back-loaded: Q4 is projected at nearly double Q1 and roughly double the mid-year trough quarters, driven almost entirely by the Holiday seasonal peak. Holiday Retail Sales — roughly 55% of the annual base — is the dominant swing factor; its forecast rests on a ~20% YoY growth rate and a seasonal shape derived from a single year of history, both of which add meaningful uncertainty. Everyday Apparel provides a stable, high-confidence counterweight at ~45% of the total, but its consistency cannot offset a meaningful Holiday miss. Leadership should pressure-test the 19.87% Holiday growth assumption against current demand signals before treating the Q4 outlook as firm.

Corporate Overhead

FY2027 Q1–Q4 forecast

Office & Admin Overhead

Corporate

Assumption-driven

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$52K	$53K	$52.5K	$54K

Projection — FY2027 · confidence: high

	Q1	Q2	Q3	Q4
Upside	$54.5K	$60.5K	$60.9K	$61.3K
Base	$54.5K	$61.8K	$62.2K	$62.6K
Downside	$54.7K	$63.2K	$63.7K	$64.2K

Rationale

Actuals are essentially flat, making the naive trend nearly meaningless as a signal. The dominant driver is a contractually signed lease renewal with a 22% rent increase effective Q2 FY2027 — the base case holds Q1 flat, then steps up ~13–14% from Q2 onward as the lease step-up is applied against total overhead.

Automated message · Slack/email to stakeholder

Forecast → Corporate leadauto-drafted

Office & Admin Overhead is projected to hold flat in Q1 FY2027 (~$54.5K) before stepping up roughly 13–14% in Q2 (~$61.8K) when the signed lease renewal takes effect with its 22% rent increase. The lease contract gives us high confidence in the timing and magnitude of this step-change. Please factor this into Q2 budget planning and any space-utilization decisions.

Employee Benefits & Insurance

Corporate

Assumption-driven

Actual — FY2026

	Q1	Q2	Q3	Q4
Actual	$98K	$101K	$103K	$104.5K

Projection — FY2027 · confidence: medium

	Q1	Q2	Q3	Q4
Upside	$113.5K	$115.7K	$118K	$120.4K
Base	$117.2K	$119.5K	$121.8K	$124.2K
Downside	$120.5K	$123K	$125.5K	$128.1K

Rationale

A carrier-notified 15% premium increase at the FY2027 renewal dwarfs the ~2.16% organic trend and is the primary driver. Assuming insurance is roughly 65–70% of total benefits spend, the 15% increase adds on top of the organic drift from Q1 onward; the upside/downside band reflects uncertainty in that assumed insurance share.

Automated message · Slack/email to stakeholder

Forecast → Corporate leadauto-drafted

Employee Benefits & Insurance is projected to jump to approximately $117.2K in Q1 FY2027 — meaningfully above the naive trend baseline — driven by the confirmed 15% insurance premium increase effective this renewal cycle. The historical ~2% quarterly organic drift continues from there, reaching roughly $124.2K by Q4. Confidence is medium given that the exact insurance cost share and renewal effective date within the year carry some model uncertainty.

Annual roll-up: Corporate Overhead

FY2027 annual forecast

Confidence: medium

Consolidated across both line items above — quarterly totals computed deterministically, confidence and commentary judged by Claude.

Quarterly base-case total — FY2027 (sum of both line items)

	Q1	Q2	Q3	Q4
Office & Admin Overhead	$54.5K	$61.8K	$62.2K	$62.6K
Employee Benefits & Insurance	$117.2K	$119.5K	$121.8K	$124.2K
Total (base case)	$171.7K	$181.3K	$184K	$186.8K

FY2027 annual base: $723.8K

Upside: $704.8K

Downside: $742.9K

Key assumptions

The Q2-effective signed lease renewal locks in a 22% rent step-up, the primary contractually certain cost driver for Office & Admin through the rest of FY2027.
Insurance premiums are assumed to represent ~65–70% of total Employee Benefits spend, making the 15% carrier-notified premium increase the dominant cost driver for the larger line item.
The insurance renewal is assumed to take full effect at the start of Q1 FY2027; a mid-quarter or delayed effective date would shift the timing of that step-change.
All non-lease and non-insurance components grow at modest organic rates (~1–2% per quarter) with no additional structural changes modeled.

Corporate Overhead in FY2027 is a rising-cost story driven by two externally-imposed step-changes — a signed lease renewal and a carrier-notified insurance premium increase — rather than organic drift, and both hit early in the year. The lease is the cleaner of the two risks: contractually locked, well-quantified, and certain from Q2 onward. The bigger swing factor is the insurance premium increase, which carries meaningful uncertainty around the assumed insurance share of total benefits spend — that single internal assumption is what separates base from downside. Leadership should validate the benefits cost composition and confirm the exact renewal effective date, as both directly determine whether the portfolio tracks base or drifts toward downside.

Two archives: the detail, and the roll-up

Every line item's forecast lands in one Data Table; every group's consolidated annual rollup lands in a second one. Both are lightweight databases built directly into this n8n workspace, not separate external systems — viewable in n8n's own Data Tables screen, reachable by any other workflow in the same project, and never exposed to the public internet on their own.

budget_forecast_projections — one row per line item, per forecast run.

forecast group	line item	method	confidence	base case (Q1 → Q4)
SaaS Revenue Streams	Recurring Subscription Revenue	assumption-driven	medium	$534.6K → $673.9K
SaaS Revenue Streams	Professional Services Revenue	trend extrapolation	high	$86K → $89.2K
Retail Sales Lines	Holiday Retail Sales	seasonal adjustment	medium	$215.9K → $371.7K
Retail Sales Lines	Everyday Apparel Sales	trend extrapolation	high	$157.6K → $172.2K
Corporate Overhead	Office & Admin Overhead	assumption-driven	high	$54.5K → $62.6K
Corporate Overhead	Employee Benefits & Insurance	assumption-driven	medium	$117.2K → $124.2K

budget_forecast_annual_rollup — one row per forecast group, per run.

forecast group	confidence	annual base	annual upside	annual downside
SaaS Revenue Streams	medium	$2.76M	$2.95M	$2.58M
Retail Sales Lines	medium	$1.47M	$1.57M	$1.36M
Corporate Overhead	medium	$723.8K	$704.8K	$742.9K

Going live: what changes outside this demo

Nothing here runs on a manual trigger in production. Historical actuals have to reach this pipeline on a schedule, and both the per-item forecasts and the consolidated annual view have to land somewhere planning actually works from.

where the history already is

General ledger or FP&A tool

QuickBooks, NetSuite, or an FP&A platform's API/export pushes trailing quarterly actuals per line item on a schedule instead of a manually pasted-in payload.

→

unchanged from this demo

This projections & roll-up pipeline

Same per-item baseline → judge → draft sequence, then the deterministic sum and consolidated commentary, running on real line items and a real fiscal year instead of synthetic ones.

→

write-back

Written back to the planning tool

Both the per-item bands and the consolidated annual rollup get written back as a forecast draft — so the next planning cycle starts from an informed number at every level, not a blank column.

→

human in the loop

Real Slack/email + a reviewed draft

High-confidence trend lines can post directly once trust is established. Assumption-driven forecasts and every annual roll-up land with the FP&A lead for a quick sanity check before a stakeholder or leadership sees it.

Why not just use a rolling forecast spreadsheet or the planning tool's built-in projections?

Adaptive Insights, Planful, Vena, and a disciplined rolling-forecast spreadsheet already extrapolate trend lines, support scenario planning, and roll line items up into a total. At a company running one of those well, this workflow isn't a replacement for it.

Where a custom pipeline like this earns its place: native trend tools apply one growth-rate formula uniformly and leave it to a human to notice when that formula is wrong for a given line, to write up the reasoning per item, and then to separately synthesize a dozen line items into one narrative a leader will actually read. This build automates all three steps — the per-item judgment call, the deterministic roll-up, and the consolidated commentary — without migrating your planning data into a new platform.