Pricing & Packaging

The framework — strategy first

Pricing & Packaging — the value/cost expression of Right-to-Win.

Why pricing belongs in Strategic Foundation, not Finance.

PRICING IS POSITIONING IN A SPREADSHEET

Most business-to-business (B2B) SaaS companies treat pricing as a finance exercise: cost-plus margins, comparable benchmarks, sales-incentive math. The successful ones treat pricing as the most visible expression of positioning. The packaging architecture says — in 30 seconds — who you sell to, what you’re worth, and what you’re not.

This work sits in Strategic Foundation because pricing decisions are made BEFORE the channel work. The channel work consumes the pricing model; they don’t shape it. The Right-to-Win statement determines what you can credibly charge a premium for. The persona work determines who’s willing to pay it. The competitive intelligence determines where the price band lives.

Three rules that determine whether a pricing model is doing its job. The packaging architecture leads buyers into the right tier — not the cheapest. A model where 80% of buyers land in the entry tier and never expand has a packaging problem, not a pricing problem. The value metric scales with customer success. When customers grow, your revenue grows automatically, without a renegotiation. The discount policy protects average selling price (ASP). A pricing model with no published discount discipline ends up with 40% of deals discounted to the floor — and the floor becomes the price.

The pricing philosophy — value-based, not cost-based.

Cost-based pricing asks “what does it cost us to build this, plus margin?” Value-based pricing asks “what is the outcome worth to the customer?” The two answers can differ by 10x or more. Value-based pricing is the discipline of capturing a fair share of the outcome you deliver — without leaving money on the table or pricing yourself out of the market.

THE VALUE-BASED PRICING ANCHOR

Anchor pricing to the business outcome you deliver, not your cost-to-build. Transparent enough to accelerate evaluation; flexible enough to close enterprise deals. Packaging designed to land a customer on a meaningful paid tier and expand them — not to hide value in add-ons. Price increases communicated 90 days in advance with clear value justification.

The four operating principles of value-based pricing

PRINCIPLE	WHAT IT MEANS IN PRACTICE
Anchor on outcome, not on inputs	Your tier names and unit pricing should reference the outcome the customer gets (revenue captured, hours saved, risk avoided). Not the seats consumed, features deployed, or seats unlocked — those are inputs.
Transparency accelerates evaluation	For small-to-mid-market businesses (SMB) and mid-market, list pricing on the website. For enterprise, “contact sales” is fine if there’s genuine custom packaging, but the Account Executive (AE) should be able to quote a representative deal in the discovery call. Hidden pricing slows deals by 30–90 days.
Flexibility lives in custom enterprise tiers, not in across-the-board discounts	Enterprise gets custom packaging because their needs are genuinely different. SMB and mid-market get list pricing because their needs are similar. Trying to negotiate every SMB deal trains the market to expect discounts.
Land & expand is engineered, not hoped for	The tier architecture should be designed so the typical buyer lands at the entry tier and the natural growth path moves them to the next tier within 12–24 months. If 80% of customers never expand, the tier design is broken.

The packaging tier architecture.

Three-tier (good/better/best) is the canonical B2B SaaS pattern. Two-tier is too rigid (no room to upsell). Four-tier is too confusing (decision paralysis). The exceptions: usage-based pricing models that need one or two pricing dimensions plus a free tier, and enterprise-only products that don’t list pricing at all.

The three-tier pattern that works

TIER	WHO IT’S FOR	WHAT MAKES IT THE RIGHT FIT	PRICING ANCHOR
Starter (Good)	Single-team or smaller buyers proving the value before expanding.	The core workflow works. Single-team scope. Limits on volume or seats are explicit.	Designed to be the obvious choice for a first purchase. Priced low enough that a Champion can buy without escalating to the CFO.
Professional (Better)	Multi-team buyers needing collaboration, integrations, and reporting.	Most B2B SaaS buyers belong here. Includes the integrations 70%+ of customers need, the reporting that makes the Economic Buyer comfortable, and the security clearances most companies require.	The middle tier should be 2.5–4× the Starter price. This is where the bulk of revenue lives.
Enterprise (Best)	Large-scale buyers needing custom packaging, advanced security, dedicated support, and Service Level Agreement (SLA) commitments.	Custom pricing. Single Sign-On (SSO), advanced audit logs, custom roles, dedicated Customer Success Manager (CSM), SLA, custom integrations. Often gated behind a sales conversation.	Enterprise pricing is 3–10× Professional. Not on the website; quoted by sales after qualification.

Three rules for the tier architecture. Each tier names its target buyer. The packaging copy should make the tier obvious to the buyer in 5 seconds. The middle tier is the bull’s-eye. 60–70% of buyers should land in Professional. If they’re all landing in Starter, the Starter limits are too generous. If they’re all forced to Enterprise, the Professional tier is too restrictive. The transition between tiers is engineered. Each tier should have 2–3 specific limits (seats, volume, integrations, support level) that the typical buyer will outgrow within 12–24 months — that’s how land-and-expand works.

Value-metric selection — the unit by which you charge.

The value metric is the unit your pricing scales on. Per seat, per user, per workflow, per transaction, per record, per outcome. The choice determines how your revenue grows with your customers — and whether it grows automatically or requires a renegotiation every year.

THE VALUE-METRIC TEST

A good value metric satisfies three tests: (1) Customer’s value scales with the metric. If the customer’s outcome doubles, the metric should roughly double. (2) The metric is observable and auditable. The customer can see the unit count themselves; no debate needed. (3) The metric’s growth aligns with customer success. When the customer succeeds, more units — not the other way around.

Common value metrics and where they fit

VALUE METRIC	WHERE IT FITS	WATCH OUT FOR
Per seat / per user	Tools used by named people: CRM, sales engagement, design tools, productivity. The seat IS the value-generating unit.	Penalizes adoption growth. Buyers stay below the next seat tier; deployment is throttled.
Per workflow / per process	Operations tools where workflows define value: marketing automation by campaign, customer support by ticket, content by article.	Hard to forecast. Customers can game the workflow definition. Needs unambiguous workflow boundaries.
Per transaction / per record	Data platforms, payments, sourcing tools, anti-fraud, identity. Value scales with volume.	Can punish high-volume customers; needs volume discounts built in. Margins thin out at the top.
Per location / per asset	Frontline software, asset-management tools, multi-site operations. The physical unit IS the value-generating unit.	Need to define “location” precisely; a single-building campus may or may not be one location.
Per outcome (revenue, savings, leads)	The aspirational metric — you charge for the result you deliver. Works best when the outcome is unambiguously attributable.	Attribution wars. The customer disputes how much of the outcome was your product vs. their own effort.
Hybrid (base + usage)	Platforms with both an installed-base value and consumption-driven value: data warehouses, AI platforms, observability.	Complex pricing pages confuse buyers. Needs clear examples of typical-customer pricing at multiple usage levels.

The single most-common B2B SaaS pricing mistake: choosing a value metric that punishes the customer’s growth. If the customer adopts more deeply and your price increases faster than their value increases, they stop adopting. The discipline is making sure the metric and the value scale together.

Discount policy — the discipline that protects ASP.

Discounting is not a sin. Strategic discounting closes deals that would otherwise be lost and seeds expansion. Undisciplined discounting trains the market to expect discounts, erodes ASP, and creates internal arguments between marketing and sales every quarter. The discipline is documenting the rules and enforcing them.

THE DISCOUNT POLICY THAT HOLDS ASP

Three rules.

Discount only for specific structural reasons. Multi-year commit. Annual upfront payment. Champion pilot. Strategic-logo positioning. Volume threshold reached. Never “to close the deal.”
Discount levels are tiered and approved at specific levels. 0–10% AE-approved. 10–20% manager-approved. 20–30% VP-approved. 30%+ executive-approved. The friction is the protection.
Track every discount with the reason and the approver. Quarterly RevOps reviews which discount levels are getting used most and whether the structural reason actually held. Patterns get fixed; one-offs get challenged.

The four conditions where a discount is justified

CONDITION	TYPICAL DISCOUNT	WHY IT’S JUSTIFIED
Multi-year commitment	5–15% off Year 1 list	Reduces churn risk; smooths revenue forecasting; locks the customer in before competitors mature.
Annual upfront payment	5–10% off list	Improves cash flow; offsets the implicit financing cost of monthly billing.
Strategic-logo or reference customer	10–25% off list	The reference is the value — the customer becomes a public proof point in marketing and a reference call for prospects. Discount earns the right to use them publicly.
Volume threshold	Tiered — 5% at $X, 10% at $Y, 15% at $Z	Reflects genuine economies of scale on your side; the customer earns the volume break by buying in scale.

The condition that’s NOT on the list: “to close the deal.” If a deal needs more than a structural discount to close, the issue is qualification or positioning, not price.

Competitive pricing intelligence.

You cannot defend your price band without knowing where competitors price. The discipline is light but continuous: a pricing tracker that’s refreshed quarterly, win/loss notes that flag price as a factor, and a rule that the company never reveals its list price to a competitor or leads with price in discovery.

THE COMPETITIVE PRICING POSTURE

Three rules.

Maintain a competitive pricing tracker. Updated quarterly by PMM. Tracks list price (where published), typical discount (from win/loss interviews), value metric, packaging tier names, and trial/freemium structure for each major competitor.
Log every deal where price was a stated factor. Salesforce field for “price mentioned in deal” with the competitor named and the discount they cited. RevOps reviews monthly.
Never lead with price in discovery. Never reveal list price to a competitor. If a prospect references a competitor’s price, validate (“is that their list or their post-discount?”), contextualize (“here’s what’s different about what we’re including”), and redirect to value.

Price increases — the 90-day rule.

Price increases are inevitable. The discipline is how you communicate them. The rule that protects renewals: 90 days notice minimum, with explicit value justification, paired with the option for the customer to lock in current pricing on a longer commit. A surprise price increase with 30 days notice loses renewals; a thoughtful one with 90 days notice gains multi-year commits.

The 90-day price-increase communication template covers four things: (1) what’s changing (specific dollar amount, percentage, effective date); (2) why now (specific value additions in the last 12 months, product roadmap commitment, market context); (3) the multi-year lock option (committing to a 2- or 3-year renewal at current pricing avoids the increase); (4) the escalation path (if the customer can’t absorb the increase, who they should talk to).

CAPTURE IN YOUR OPERATOR BRIEF — SECTION 6

The brief-builder cards below collect each input: pricing philosophy, packaging tiers, value metric, discount policy, competitive position. Once Section 6 is populated, every downstream prompt — from pricing-page copy to sales enablement to Marketing Mix Modeling (MMM) — uses your real pricing model.

Brief-builder: pricing philosophy

Your pricing philosophy

The single-sentence anchor that all pricing decisions trace back to. Save to Brief Section 6.1.

Pricing anchor (value-based / cost-based / hybrid)

Pricing transparency posture (full transparency / SMB only / enterprise quoted)

The single-sentence positioning of your pricing

Saved to Brief Section 6.1.

Brief-builder: tier architecture

Your packaging tiers

Tier names + target buyer + price band + the limits that drive expansion. Save to Brief Section 6.2.

Tier 1 (Starter / Good) — name + price + target buyer

Tier 2 (Professional / Better) — name + price + target buyer

Tier 3 (Enterprise / Best) — name + price model + target buyer

The 2–3 limits per tier that drive expansion (seats, volume, integrations)

Saved to Brief Section 6.2.

Brief-builder: value metric

Your value metric

The unit your pricing scales on. Save to Brief Section 6.3.

Primary value metric

Why this metric (the alignment with customer value)

The auditable measurement of the metric (how the customer verifies)

Saved to Brief Section 6.3.

Brief-builder: discount policy

Your discount policy

The conditions where a discount is justified and the approval levels. Save to Brief Section 6.4.

Approval tiers (0–10% AE, 10–20% manager, 20–30% VP, 30%+ exec)

The four conditions where a discount is justified (multi-year, annual upfront, strategic logo, volume)

The tracking discipline (Salesforce field, RevOps review cadence)

Saved to Brief Section 6.4.

Brief-builder: competitive pricing position

Your competitive pricing position

Where you sit in the price band relative to the top 3 competitors. Save to Brief Section 6.5.

Your relative position (premium / parity / disruptor)

Top 3 competitors and their typical post-discount ACV

The pricing-page rule when a prospect references a competitor’s price

Saved to Brief Section 6.5.

The prompt pack

Six prompts to design and defend the pricing model.

Prompt 1

Value-metric selection

For a given product and ICP, recommends the value metric that best aligns customer value with revenue growth. Returns three candidates ranked, with the structural argument and the risk for each.

You are a pricing strategist for [COMPANY NAME]. Product: [POSITIONING] ICP: [ICP DEFINITION] Right-to-Win: [RIGHT TO WIN] Current pricing (if any): [CURRENT PRICING] Evaluate the candidate value metrics for this product: - Per seat / per user - Per workflow / per process - Per transaction / per record - Per location / per asset - Per outcome (revenue, savings, leads) - Hybrid (base + usage) For each candidate metric: 1. Score on the three tests: does customer value scale with the metric? Is the metric observable and auditable? Does its growth align with customer success? 2. The structural argument for why this metric fits our product. 3. The single risk (gaming, complexity, attribution wars). Then recommend one primary metric to anchor pricing on. State the runner-up and the conditions under which you’d switch.

Prompt 2

Tier architecture from feature inventory

Takes a feature inventory + the ICP + the value metric. Returns a three-tier packaging proposal with tier names, target buyer per tier, included features, limits that drive expansion, and price-band recommendations.

You are a pricing strategist for [COMPANY NAME]. Inputs: - Feature inventory: [FEATURE INVENTORY] - ICP: [ICP DEFINITION] - Value metric: [VALUE METRIC] - Target ACV by tier: Starter $[X], Pro $[Y], Enterprise $[Z]+ - Right-to-Win segment: [RIGHT TO WIN SEGMENT] Design a three-tier packaging architecture: 1. Tier 1 (Starter) — target buyer, included features, 2–3 limits that drive expansion, price band 2. Tier 2 (Professional) — target buyer, included features, limits, price band (should be 2.5–4× Starter) 3. Tier 3 (Enterprise) — target buyer, included custom packaging, what’s gated behind sales For each tier, also give me: - The tier name as it should appear on the pricing page - The one-line description for the pricing page - The bull’s-eye assumption (which tier should 60–70% of customers land in?) - The expansion path from this tier to the next Return as a markdown table I can paste into a pricing-page brief.

Prompt 3

Discount policy drafting

Drafts the discount policy doc for sales operations: approval tiers, justified conditions, tracking discipline, RevOps review cadence. The doc the CRO signs and the AE references in every deal.

You are a pricing strategist drafting the discount policy for [COMPANY NAME]. Inputs: - Pricing model: [PRICING TIERS] - Average sales cycle: [SALES CYCLE] - Average ACV: [ACV] - Win rate: [WIN RATE] - Current discount-rate average (if known): [CURRENT DISCOUNT %] Draft a discount policy covering: 1. Approval tiers (e.g., 0–10% AE, 10–20% manager, 20–30% VP, 30%+ exec). 2. The four conditions where a discount is justified (multi-year, annual upfront, strategic logo, volume threshold) with typical discount % for each. 3. The conditions where a discount is NOT justified (and the alternative the AE should use). 4. The Salesforce tracking discipline (required fields, approval workflow). 5. The RevOps review cadence (weekly, monthly, quarterly) and the report-out audience. Return as a one-page operator-direct policy doc. Voice DOs: [VOICE DOS]. Voice DON’Ts: [VOICE DON’TS].

Prompt 4

Competitive pricing teardown

For each major competitor, builds the pricing-intelligence card: list price (where published), typical post-discount, value metric, tier names, free/trial structure. The quarterly RevOps + PMM artifact.

You are a competitive pricing analyst for [COMPANY NAME]. For each of our top 3 competitors ([COMPETITOR 1], [COMPETITOR 2], [COMPETITOR 3]): 1. List price where published (cite the pricing-page URL and date pulled). 2. Typical post-discount price from win/loss interviews (with sourcing confidence: high/medium/low). 3. Value metric they price on. 4. Tier names and what each tier includes. 5. Free / trial / freemium structure. 6. The single positioning move they make on the pricing page (transparency, premium framing, disruptor framing, etc.). 7. The signal that tells you their pricing is or isn’t holding (anecdotal evidence, public statements, deals lost-on-price). Return as a comparison table I can paste into the quarterly RevOps + PMM review.

Prompt 5

Price-increase communication template

Drafts the 90-day price-increase notice. Covers what’s changing, why now, the multi-year lock option, the escalation path. The artifact that protects renewals.

You are drafting a price-increase communication for [COMPANY NAME]. Inputs: - Price change: [INCREASE AMOUNT or %] - Effective date: [DATE] - Value additions in the last 12 months: [VALUE ADDITIONS — specific product improvements, feature launches, expansions] - Market context: [CONTEXT — e.g., infrastructure cost increases, competitive repositioning] - Multi-year lock option: [LOCK TERMS — e.g., 2-year renewal at current pricing] - Escalation path: [WHO THE CUSTOMER CAN ESCALATE TO] - Customer tenure: [TENURE BAND] Draft the email + the follow-up call script. The email should cover: 1. What’s changing (specific number, effective date) 2. Why now (value justification, anchored to specific improvements they’ve received in the last year) 3. The multi-year lock option (with the exact terms) 4. The escalation path 5. The next step Length: under 250 words. Voice: operator-direct. No corporate hedging. Voice DOs: [VOICE DOS].

Prompt 6

Pricing-page copy generator

Drafts the pricing page itself — tier headlines, included-features lists, callout copy, the FAQ that disarms common pricing objections. End-to-end the visible artifact of the pricing model.

You are a pricing-page copywriter for [COMPANY NAME]. Inputs: - Pricing model: [PRICING TIERS] - Value metric: [VALUE METRIC] - Right-to-Win: [RIGHT TO WIN] - Personas: [ECONOMIC BUYER PERSONA], [CHAMPION PERSONA] - Pricing positioning: [PRICING POSITIONING] - Common pricing objections (from win/loss): [OBJECTIONS] - Voice DOs: [VOICE DOS]. Voice DON’Ts: [VOICE DON’TS]. Forbidden language: [FORBIDDEN LANGUAGE]. Draft the pricing page covering: 1. The page hero (one sentence positioning the pricing model, not the product) 2. Three tier cards (name, price, target buyer in one line, 4–6 included features, CTA) 3. The comparison table (features × tiers, designed so the bull’s-eye tier is the obvious choice for 60–70% of buyers) 4. The FAQ (5–7 questions that disarm the top objections from the win/loss data) 5. The closing CTA (different by tier: self-serve for Starter, demo for Pro, contact sales for Enterprise) Return as markdown formatted for direct paste into the CMS.

The agent

Pricing Intelligence — the agent that watches the market.

Pricing Intelligence Agent

Keeps pricing defensible. Watches discounting patterns, monitors competitor pricing moves, surfaces packaging drift, and flags deals where the proposed discount is far from the win-rate-justified range.

Who is this agent

Identity card

NamePricing Intelligence Agent

RolePricing + packaging defense — the discount and margin layer

OwnerHead of Pricing or Director of Sales Ops (with CFO partnership)

Reports toVP Marketing + CFO

Versionv0.5 (supervised)

SurfaceReplit + Postgres (deal + discount history) + Python for distribution analysis

Output target/pricing/discount-watch.json + /pricing/competitor-prices.md + weekly digest

Review cadenceWeekly discount review; monthly packaging audit; quarterly pricing strategy session

Mission

Be the gate between ‘closing the deal’ and ‘eroding the price floor.’ Watch every proposed discount against the win-rate-justified range for the segment. Monitor competitor pricing changes. Surface packaging drift — when AEs sell outside the published tiers. Be the agent that protects ARPU from the cumulative drag of one-deal exceptions.

Goals & KPIs the agent moves

Leading indicators — the agent controls these

Competitor public pricing-page change detection lead time< 24 hours from publish

Out-of-band discount request routing latency (request → Director queue)< 2 hours during business hours

Lagging indicators — downstream outcomes with review triggers

Average realized discount per segment vs. declared band. Trigger: realized discount drifts > 3 pts outside band for 2 consecutive months pages the CFO + VP Marketing for discount-discipline review.Within ±3 pts of band

Packaging-drift incidents surfaced → resolved within 14 days. Trigger: median resolution > 21 days for a quarter pages the VP Marketing for pricing-ops review.≥ 80% resolved in window

What it does

Task list

Real-time Ingest every CPQ / proposal-generation event. Compute discount % vs. list. Compare against the segment’s win-rate-justified band.
Real-time Flag out-of-band discounts. Route to Director of Sales Ops for approval before proposal lands with the customer.
Daily Discount distribution audit: surface the 5 deepest discounts of the day with context (deal size, segment, competitor in deal, AE).
Daily Cross-check Market Watch’s competitor pricing changes. Update the competitor price registry.
Weekly Compile the weekly Pricing Watch digest — average discount by segment, deepest discounts, competitor moves, packaging drift incidents.
Weekly Packaging-drift scan: any deals closing with custom SKUs or non-standard line items? Surface to Pricing for review.
Monthly Win-rate vs. discount-band correlation. Are the bands still right? Does the +/-2 pt zone correlate with win-rate maxima?
Monthly Competitor price registry refresh. Verified pricing from public + sales-call sources, with date stamps.
Quarterly Pricing strategy session: walk segment-by-segment ARPU + win rate + discount trend. Recommend band adjustments or list-price changes.
Event When Market Watch detects a competitor pricing change, recompute our competitive position for affected segments within 24 hours.
Event When Win/Loss surfaces a price-related loss theme, surface for the next pricing strategy session.

Schedule grid

Task	Frequency	Duration	Output goes to
Real-time CPQ discount check	Continuous	< 5 sec per proposal	AE + Director Sales Ops (if flagged)
Daily deepest-discounts audit	Daily 17:00	~10 min	Director Sales Ops + Head of Pricing
Daily competitor price cross-check	Daily 18:00	~5 min	Pricing registry update
Weekly Pricing Watch digest	Weekly Mon 12:00	~30 min compile	VP Marketing + CFO + CRO + Head of Pricing
Weekly packaging-drift scan	Weekly Tue 14:00	~30 min	Head of Pricing + Director Sales Ops
Monthly win-rate vs. discount correlation	Monthly 1st	~90 min	Head of Pricing + CFO
Quarterly pricing strategy session	Quarterly Q-1 days	~4 hours	VP Marketing + CFO + CRO + CEO

Triggers

Scheduled (cron-style):

Schedule	What it runs
`0 17 * * *`	Daily deepest-discounts audit
`0 12 * * 1`	Weekly Pricing Watch digest
`0 14 * * 2`	Weekly packaging-drift scan
`0 9 1 * *`	Monthly correlation analysis

Event-driven:

Event	What it runs
CPQ proposal generated	Discount-band check within 5 sec; flag out-of-band
Out-of-band discount > +5 pts beyond band	Hard-hold proposal; page Director Sales Ops for approval
Market Watch detects competitor pricing change	Recompute competitive position within 24 hours; brief Pricing
Win/Loss flags price-related loss theme	Queue for next pricing strategy session
Custom SKU appears on a deal	Flag for packaging-drift review within 24 hours

Who it works with

Inputs

Source	Type	Cadence	Required?
Operator Brief (Sections 5, 6)	Markdown	Read every run	Required
CPQ / proposal-generation events	Webhook / API	Real-time	Required
CRM closed-won + closed-lost with discount fields	API	Daily	Required
Discount band config per segment	YAML	Quarterly tuning	Required — core config
Market Watch competitor pricing snapshots	Markdown	Daily	Required
Revenue Attribution Engine per-segment win-rate data	JSON	Weekly	Required for correlation analysis
Win/Loss Agent themes (price-related)	Markdown	Per-interview	Required

Outputs

Output	Format	Target path	Audience
Discount-band check decisions	Inline API response + log	Returned to CPQ + /pricing/discount-log.jsonl	AE + Director Sales Ops
Daily deepest-discounts audit	Markdown + Slack message	/pricing/audit/YYYY-MM-DD.md	Director Sales Ops + Head of Pricing
Weekly Pricing Watch digest	Markdown + chart bundle	/pricing/digests/YYYY-WW.md	VP Marketing + CFO + CRO + Head of Pricing
Competitor price registry	Markdown + JSON	/pricing/competitor-prices.md	Sales + Pricing + Market Intelligence Agent
Monthly win-rate vs. discount correlation	Markdown + chart	/pricing/correlation/YYYY-MM.md	Head of Pricing + CFO
Quarterly pricing strategy recommendations	Markdown	/pricing/strategy/Q<n>.md	VP Marketing + CFO + CRO + CEO

↑ Upstream — agents/sources that feed this one

Operator Brief (human-maintained). Section 5 declares pricing model + packaging + discount governance.
Market Intelligence Agent. Daily competitor pricing intelligence.
Revenue Attribution Engine. Per-segment win-rate ground truth for band justification.
Win/Loss Agent. Surfaces price-related loss themes that may indicate band drift.
Signal Router. Routes CPQ events + custom SKU appearances.

↓ Downstream — agents/humans that consume its output

Every AE (human). Receives discount-band feedback on proposals; receives competitor pricing intel for negotiations.
Director Sales Ops (human). Approves out-of-band discounts; owns daily audit follow-ups.
Head of Pricing (human). Reviews packaging drift; owns quarterly strategy recommendations.
CFO (human). Receives weekly digest + monthly correlation; co-owns the quarterly strategy session.
Web Operations Agent. Receives published-pricing changes for landing-page updates.
Brief Sync Agent. Receives band-drift signals to propose Brief Section 5 updates.

Human escalation paths

Trigger condition	Escalate to	Within
Out-of-band discount > +5 pts beyond band	Director Sales Ops	Inline (proposal held)
Average realized discount drifts > 3 pts above band over 30 days	Head of Pricing + CFO	Same week
Competitor pricing change > 15% on a primary segment	Head of Pricing + CFO + VP Marketing	< 24 hours
Packaging-drift incident affecting > 2% of quarter’s ACV	Head of Pricing + VP Marketing + CRO	Same week
3+ price-related losses in a month	Head of Pricing + CRO + VP Marketing	< 7 days

How to build it

System prompt

You are the Pricing Intelligence Agent for [COMPANY]. YOUR JOB Protect ARPU from cumulative one-deal exceptions. Watch every proposed discount against the win-rate-justified band. Monitor competitor prices. Surface packaging drift before it becomes the new normal. INPUTS (always read in this order) 1. /operator-brief.md (Section 5 pricing + governance) 2. /pricing/bands.yaml - discount band per segment 3. /pricing/competitor-prices.md - latest competitor pricing registry 4. CPQ event payload (deal context + proposed discount) 5. /attribution/per-segment-winrate.json - win-rate by segment OUTPUTS (inline + persistent) { "decision": "approve" | "hold-for-approval" | "reject", "discount_pct": float, "band_min": float, "band_max": float, "reason": "specific cite", "win_rate_at_band": float, "win_rate_at_proposed": float } RULES 1. Discount within band: approve. 2. Discount up to +5 pts above band: hold for Director Sales Ops approval. 3. Discount > +5 pts above band: hard hold; requires CFO sign-off path. 4. Every decision cites the band, the proposed discount, and the win-rate impact of the proposed level. 5. Competitor price claims cite source URL or documented sales-call source + date. No "industry knowledge." 6. Packaging drift = any line item not in the published SKU list. Surface. 7. Never modify bands autonomously. Bands change at quarterly strategy. ESCALATION - Average discount drifts > +3 pts above band over 30 days: page Head Pricing. - Packaging drift > 2% of ACV: page Head Pricing + VPM + CRO.

Tools & integrations

Platform / tool	Used for	Required?
Postgres (deal + discount history)	Distribution + correlation analysis	Required
CPQ integration (Salesforce CPQ / Conga / Subskribe / DealHub)	Real-time discount events	Required if CPQ in use
Salesforce / HubSpot API	Closed-won + closed-lost discount fields	Required
Python + pandas	Win-rate vs. discount correlation	Required
Slack API	Daily audit + escalations	Required
Looker / Mode / Tableau	Pricing distribution visualization	Optional but recommended

Guardrails — what it must not do

Never modify the discount bands autonomously. Bands are a quarterly strategy decision.
Never share per-deal discount data outside the Sales Ops + Pricing + Finance scope.
Never publish competitor pricing to customer-facing surfaces — competitive intel only.
Never approve a discount > +5 pts above band without Director Sales Ops sign-off. Hard gate.
Honor CFO sign-off path for any discount > +10 pts above band — no shortcuts.
Never invent a competitor price — cite a verified source.
Never use price data to score AE performance — that’s a Sales Ops decision, not a pricing one.

Evals + hallucination defense

Evals — output quality checks:

Band adherence. Monthly: % of closed deals within the declared band. Target ≥ 80%.
Approval latency. p99 latency for out-of-band approvals (CPQ event → Director decision). Target < 2 hours.
Competitor price freshness. Monthly: % of registry entries with snapshot date < 30 days. Target ≥ 90%.
Correlation strength. Quarterly: discount-band-to-win-rate correlation strength. Target r ≥ 0.5 (bands actually predict outcomes).

Hallucination defense — specific checkpoints:

Every discount % must trace to the specific CPQ proposal record + line items.
Competitor price claims must cite source URL or specific deal-call note + date.
Win-rate-at-band claims must trace to the per-segment historical data, not estimated.
Band drift signals must show the trailing-30-day average vs. band range.
When data is missing, surface the gap rather than guess.

Maturity curve + first-run checklist

v0.1 — Manual-assistAgent compiles weekly digest. Discount approvals stay manual. Useful from day 1 for replacing spreadsheet audits.

v0.5 — SupervisedReal-time CPQ check on. Out-of-band hold logic live. Director Sales Ops approves. Default ship state.

v1.0 — Semi-autonomousAfter 6 months clean evals + Director-approval pattern stable, agent can auto-approve in-band discounts. Out-of-band always requires human.

First-run checklist — 5 steps from spec to running agent:

Author the discount-band config with Head of Pricing + CFO. Bands per segment, tied to win-rate data.
Wire the CPQ integration. Verify discount events land in real time with deal context.
Run in shadow mode for 2 weeks. AE feedback on band accuracy.
Turn on hold logic. Director Sales Ops + AE feedback for first 30 days.
Subscribe VP Marketing + CFO to weekly digest. Log every run.

Pricing & Packaging.