Home
Features
Financial Intelligence

Financial Intelligence · The quiet engine

The money side of the business, watching itself

Six tools that catch margin gaps before you send the estimate, flag cost overruns the same day they happen, and build next year's P&L for you. No spreadsheets. No month-end surprises.

Book a Demo See Pricing

AI estimate scrubber Live margin badges Real-time job costing Spend anomaly alerts AI upsells with rationale Annual budget architect

Estimates · live margin

Hartwell HOA — spring mulch & cleanup $8,420 Good · 41%

Brookside corporate — turf renovation $14,180 Low · 28%

Maple Lane — emergency tree removal $3,260 Negative · −6%

Why operators turn this on

Most landscape companies find the loss eight weeks too late.

The bid went out underpriced. The crew burned an extra day on materials. A vendor crept up 30% over the quarter. None of it surfaces until you close the books — and by then the season is gone.

The bid that lost money before it shipped

Estimate goes out at 18% margin because the labor line was light and nobody caught it. The signed contract now binds you to a job that breaks even at best. Found at month-end — too late to renegotiate.

The job that quietly slid into the red

Crew's been at it three days longer than estimated. Materials reorders ran 22% over. The PM sees it on a printed report Friday morning — after Wednesday's losses are already booked.

The vendor charge nobody noticed

Same supplier you always use. New invoice rate creeps up 18%. You catch it on the credit-card reconciliation 47 days later. Multiply by every vendor you have, and that's the margin you can't find.

How it works

From estimate to year-end, watched all the way through.

One pipeline. The estimate gets scrubbed before you can send it; the job gets costed while it's live; the spend gets watched in the background; the year gets planned from a real P&L. Each stage feeds the next.

Estimate drafted — the scrubber runs before send

You finish a draft. RootControl runs it through deterministic rule checks: negative margin, low margin against your target threshold, missing labor lines, individual lines priced below cost. Issues come back tagged error, warning, or info, each pointing at the specific line that triggered it. Fix in seconds, or override with a reason on file.

AI suggests the add-ons you'd otherwise forget

One tap on the draft asks the AI for upsells. It looks at what's on the estimate, your service catalog, and known complementary pairings — pre-emergent with mulch, soil amendments with planting, edging with bed installs — and returns a short list with a one-sentence written rationale for each. Accept or skip, then send.

Job goes live — costing tracks actual vs. estimated

Once the work starts, every timesheet punch, material entry, and subcontractor charge rolls up against the estimate snapshot in real time. Labor, materials, and equipment each track separately, with approved and pending actuals shown side by side. Jobs flip from on target to at risk to over budget at a 10% variance threshold — with the line items that caused the slip linked one click away.

The background watches spend — and the year plans itself

Spend Anomaly Detection runs nightly against a rolling 90-day baseline per vendor and category. Anything that crosses ~1.5σ (or 1.25× the mean when data is thin) fires an alert with a one-sentence AI summary and a suggested action. Meanwhile, the Operating Budget Architect builds your annual P&L from industry benchmarks calibrated to your revenue — monthly variance tracking flips categories to at risk at 90% of budget and over at 100%.

What's in the suite

Six tools, one financial nervous system

Each tool earns its keep on its own. Together they catch the leaks at every stage — before send, during the job, across the quarter, and over the year.

AI Estimate Scrubber

Every draft runs through deterministic rule checks before it can be sent. Margin gaps and coverage holes are caught at the source — not at month-end.

Severity tagged per issue: error, warning, or info
Margin target defaults to 35% — configurable per org
Rules covered: negative margin, low margin, missing labor, lines priced below cost
Each issue links to the line that triggered it — fix in seconds
Override with a reason on file when a strategic bid undercuts the threshold

All plans

AI Upsell Suggestions

A short list of complementary add-ons on every estimate — each with a one-sentence written rationale so you (and the customer) know why.

Suggestions pulled from your service catalog plus known complementary pairings
Worked examples: pre-emergent with mulch, soil amendments with planting, edging with bed installs
Each suggestion arrives with a rationale string — written, not implied
Added as optional addon lines you accept or reject before send
Acceptance rate tracked so you can see which pairings actually land

All plans

Live Margin Badges

Three states, glanceable from any estimate list. Green for Good, amber for Low, red for Negative — the badge updates as you edit, so you never send blind.

Good / Low / Negative status computed from the same threshold the scrubber uses
Visible on the estimate detail and on every list view
Margin and revenue totals shown side-by-side with the badge
Filter the estimate queue by badge to triage the weakest bids first
Recalculates instantly when you add, remove, or reprice any line

All plans

Real-Time Job Costing

Actual vs. estimated, while the job is still in flight. Labor, materials, and equipment tracked separately — with approved and pending actuals shown side by side.

Estimate snapshot locked at job start so the comparison stays clean
Per-category breakdown: labor, materials, equipment, subcontracting
Approved vs. total actuals so unbilled cost doesn't hide in pending
Status: on_target / at_risk / over_budget at a 10% variance threshold
Drill from the status badge to the specific line that caused the slip

Professional+

Spend Anomaly Detection

Background watcher on every vendor and category. Cost overruns, runaway fuel, vendor-rate creep — flagged the day they happen, not at month-end close.

Rolling 90-day baseline per vendor & per category
Alert fires at ~1.5σ over baseline (or 1.25× mean when data is thin)
Severity tagged: info, warning, or critical
14-day dedup window — no spam from the same category
AI drafts a one-sentence summary plus a concrete suggested action on every alert
Daily anomaly digest to the inbox; one-click escalate to manager review

Professional+

Operating Budget Architect

A starter annual P&L generated from industry benchmarks calibrated to your revenue. Edit, lock, and watch the variance run live against actuals all year.

Benchmark allocations: ~30% field labor, ~20% materials, ~8% equipment, ~22% overhead, ~15% target profit, ~5% reserves
Monthly P&L bucketed by fiscal year, drillable to category
Variance tracking: at_risk at 90% of budget, over_budget at 100%+
Separate operating and overhead budgets (field labor vs. office labor, etc.)
Actuals stream in from invoices, timesheets, and material/PO entries — no manual entry
3-month AI seasonal revenue forecast layered on top of the plan

Professional+

Under the hood

The exact thresholds we run on

No magic. The system uses deterministic rules wherever it can — and labels the AI layer clearly wherever it adds judgment. Everything below is configurable.

Defaults that ship with every org

Estimate margin target: 35% — the scrubber fires low_margin below this, negative_margin below zero. Set per org; the same threshold drives the Good / Low / Negative badge.
Scrubber rule codes: negative_margin (error), low_margin (warning), missing_labor, line_underwater. New rules can be added without code changes.
Job costing variance: 10% — jobs flip from on_target to at_risk at the threshold, and to over_budget on hard breach. Tracked per category, not just job-total.
Spend anomaly trigger: current spend > mean + 1.5σ over a rolling 90-day baseline, or > 1.25× mean when the baseline has too little history.
Anomaly dedup window: 14 days per (vendor, category) pair — one alert covers the spike, follow-ups are suppressed until the window closes.
Budget variance bands: at_risk at 90% of category budget consumed; over_budget at 100%+. Computed per category, per month.
Upsell rationale: every AI suggestion ships with a written one-sentence rationale — it's a required field on the suggestion contract, not an afterthought.

All thresholds are defaults — raise the margin target if your business runs leaner, soften the spend alert sigma if you want a louder watch, change the variance band if your contracts run tighter than the standard. Settings live with the org, not the user.

Starter annual P&L

How the Budget Architect carves up revenue

The architect drops in industry-benchmark allocations on day one so you have a real plan instead of a blank sheet. Override any line; the variance tracking runs from your edited version, not the default.

30%

Field labor

20%

Materials

Equipment

22%

Overhead

15%

Target profit

Reserves

Questions we hear

Financial Intelligence FAQ

What is the AI Estimate Scrubber, and what does it actually check?

The scrubber runs every draft estimate through deterministic rule checks before you can send it. Issues come back tagged error, warning, or info. The default rules cover negative-margin estimates, low-margin estimates under your target threshold (35% by default, configurable per org), missing labor lines, and individual lines priced below cost. Each issue points to the specific line that triggered it so you can fix it in seconds — or override with a reason on file.

Where do AI upsell suggestions come from?

When you finish a draft, you can ask RootControl to suggest add-ons. The AI looks at what's on the estimate, your service catalog, and known complementary pairings (pre-emergent with mulch, soil amendments with planting, edging with bed installs) and returns a short list of suggested lines — each with a one-sentence written rationale. Suggestions are added as optional addon lines you accept or reject before sending. Acceptance rate is tracked so you can see which pairings actually land with your customers.

How fast does Spend Anomaly Detection catch a problem?

Same day. The service runs a rolling 90-day baseline per vendor and per category. When current spend crosses ~1.5 standard deviations above the baseline (or about 1.25× the mean when there isn't enough history yet), it fires an alert tagged info, warning, or critical. A 14-day dedup window prevents the same category from spamming the inbox, and the AI layer drafts a one-sentence summary plus a concrete suggested action on every alert.

What's the difference between Real-Time Job Costing and Budget variance?

Job Costing tracks one in-flight job against its estimate — labor, materials, and equipment, with separate "approved" and "total" actuals so you can see what's been signed off vs. what's still in review. Jobs roll up to on_target, at_risk, or over_budget at a 10% variance threshold. Budget variance tracks the operating P&L: monthly category totals against the plan, with at_risk firing around 90% of budget and over_budget at 100% or more. Both run live — no month-end close required.

Does the Operating Budget Architect just make up numbers?

No. It builds a starter annual P&L using industry benchmark allocations — about 30% field labor, 20% materials, 8% equipment, 22% overhead, 15% target profit, and 5% reserves — calibrated to your historical revenue. You edit anything you want before locking it. Once locked, monthly variance tracking runs live against actuals from your invoices, timesheets, and PO/material entries. Edit again at any point in the year; the variance recomputes against the new plan from that month forward.

What plan are these features on?

The AI Estimate Scrubber, AI upsell suggestions, and Live Margin Badges ship on every plan, including Starter — they're part of how every estimate moves through the system. Real-Time Job Costing, Spend Anomaly Detection, and the Operating Budget Architect are Professional and Enterprise features. See the pricing page for the full breakdown.

Can I override the scrubber if I'm intentionally taking a low-margin bid?

Yes. Every issue can be acknowledged with a reason on file — "strategic loss-leader for a multi-year contract" or "spring discount for an anchor HOA," for example. The override is recorded on the estimate, visible in audits, and rolled up in the analytics so you can see how often your team is taking thin bids and on what justification. Strategic bids stay strategic; the rest still get caught.

Get Started

See your last quarter, scrubbed

30-minute walkthrough — bring a real estimate that lost money and a real vendor invoice that crept up. We'll show you exactly when RootControl would have flagged each one and what it would have done about it.

Book a Demo See Pricing