QuoteAI — Business Model & Engagement#

Status: 🟢 Active — persistent strategic doc Authors: Dan Hannah & Clay Created: 2026-04-22 (split from pitch.md on 2026-04-23) Parent: QuoteAI Project Design Doc | Companion: Andy Meeting Brief

Purpose#

Persistent business model and engagement framework for QuoteAI. Distinct from meeting-specific artifacts — see andy-meeting.md for tactical prep for any one meeting.

This doc evolves with the engagement over time. Meeting docs are one-off and get archived after the meeting happens.

The Model — "Buy me, not a product"#

QuoteAI is not sold as shrinkwrapped SaaS. Brehob hires Dan on a monthly retainer to deliver incremental, bespoke value against their specific workflows. The app is the living deliverable of that work.

• Why right for this stage: no team, no support org, no roadmap committee. Selling "software" would misleadingly abstract over reality.
• Why right for Brehob: de-risked buy-in. Each month is new deliverables against measurable value, not a commitment to vaporware.
• Why right for Dan: every new request becomes paid work. Scope extension IS the business model.

Product with Dedicated Engineering Support (PITCH-D2)#

The engagement is presented externally as software custom-fit to Brehob, with dedicated engineering support from Dan to keep it working and extending it as needs evolve. Product framing, with Dan-as-engineer as a feature of the product (dedicated support), not a separate billable service.

Never say "consulting" in customer-facing material — the word carries transactional / hourly / fungible connotations that undercut the pitch. That vocabulary stays with legal and accounting only.

Real-world analog: hiring an architect with ongoing involvement. You're "getting a house" (product framing — what you live with); the contract is for architectural services (structural framing — how the legal agreement flows). The house exists because of the architect's work; tightly coupled; but what Brehob experiences and talks about is the tool.

Externally (pitch, leadership, CFO):

• "You're buying software custom-fit to Brehob, with dedicated engineering support from me."
• The $5K/mo covers the tool + ongoing engineering to improve it
• Value on Brehob's P&L: "quotes generated, salespeople more productive, deals closed faster"

Internally (legal / tax / accounting):

• Structurally a Master Services Agreement (MSA) with monthly retainer terms
• Dan is a 1099 vendor; Brehob expenses the retainer as "professional services"
• The MSA is where the services framing lives; the pitch narrative never mentions it

Downstream implications resolved cleanly:

• IP: Dan owns the code (architects own their design methods); Brehob gets perpetual non-exclusive use of their specific implementation
• Renewal: Brehob renews access to the dedicated engineering support, not a software license
• Termination: they can stop paying anytime; they keep what's been built to date; no re-billing for existing capability — only for new work

If asked at the meeting "are we buying the tool or your time?":

"You're buying software that solves this specific problem for Brehob, with dedicated engineering support from me to keep it working and extending it as your needs evolve. Structurally it's a Master Services Agreement — but in practice you get a working tool plus an engineer who cares about making it better for you."

See quoteai/decisions.md § PITCH-D2 for full rationale + alternatives considered.

The Thesis — Why Bespoke Software Is Possible Now#

Ten years ago, if a company like Brehob wanted software tailored to their workflow, the options were:

• License a giant tool (Salesforce, SAP) and spend $500K+ on consultants customizing it — the tool bends you more than you bend it
• Hire an internal engineering team and spend $2M+ building from scratch
• Accept that software won't fit and keep using spreadsheets

Most industrial distributors (like Brehob) picked option 3 — the one that meets reality where it lives, but at the cost of every salesperson re-typing the same data, every admin retyping quotes, every minute spent on formatting rather than selling.

Agentic AI changes the math by ~100x. A single engineer with the right tools now builds software genuinely custom-fit to a specific workflow for the price of a mid-range SaaS subscription. The output is bespoke; the cost is commodity.

What this unlocks for Brehob specifically:

• Your existing spreadsheet is the source of truth — not a starting point for custom fields inside Salesforce
• Your 22 years of quote history is the training data — not proprietary data locked in a vendor's cloud
• Your salespeople's workflow stays intact — the tool wraps what they already do, rather than making them learn a new system
• New features land in days, not procurement cycles — because the engineer knows your business, not a generic SMB persona

This isn't "AI for sales." It's your workflow, made durable as software, by someone who knows the domain cold (John) and has built ML/AI systems in far more demanding environments (Dan — F1).

The tool is the proof of concept. The thesis is the product.

Competitive Position#

Created 2026-05-04 from research sweep — full deep dive in competitive-landscape.md.

QuoteAI is not first to category. Canals, Distro, BoltWise, and Endeavor AI all emerged 2024-2025 targeting AI quoting for industrial distributors. But the differentiation is sharp — they're solving inbound RFQ → quote automation; we're building an institutional knowledge platform with quoting as the first surface.

The defensible positioning#

Tacton codified configuration rules for manufacturers — $250K-$1M+ ACV at Caterpillar and Siemens. Canals and Distro are codifying RFQ parsing for distributors — sub-$50K ACV at Locke Supply and Puget Sound Pipe. We're codifying QUOTING EXPERTISE ITSELF — every spec call, every margin negotiation, every won/lost reason from 22 years of Brehob's history. The legacy CPQ pricing ceiling tells us the market will pay for vertical depth. The Canals/Distro round sizes tell us distributor-AI is fundable. We sit in the gap between them, $80-200K ACV.

Pricing implications — comp data validates current ranges#

Updated 2026-05-05 post-Andy meeting #2. Two distinct pricing tiers now apply:

For first-customer pilot pricing (Brehob, Year 1): see decisions.md § PRICING-D2. Andy's read landed at:

• Flat monthly: ~$3,500/mo
• One-time setup/implementation: $5-10K
• Year 1 target: ~$47-52K

This is intentionally below the comp range — first-customer "price to get in" anchor. Andy's framing: "Get it in there, get it, you know, let people get it in, break it, figure it out, learn from it." Multi-phase pricing growth (each new scope addition is a new pricing conversation) brings Year 2+ pricing into the comp range as scope expands.

For scaled-customer comp range (Brehob Year 2+, Customer #2 onward):

• Sales enablement (most relevant comp by buyer): Highspot $91K avg ACV at our scale (verified 62-deal Vendr sample); category band $80-130K
• Vertical CPQ for our class: $75-200K ACV
• Distributor-AI startups (Canals, Distro): sub-$50K likely
• Defensible QuoteAI scaled range: $80-200K ACV

The earlier consideration to raise Option C base $5-7K → $7-10K/mo is superseded. Andy's read pointed downward for the pilot, not upward. Highspot's $91K is the Year-2+ ceiling target reached through scope expansion — it's not a Year-1 floor.

Implication for current pricing options menu (## Pricing Strategy § Two Pricing Options):

• The Option A/B/C/D menu is now reserved for Year-2+ Brehob and future-customer reasoning
• Brehob Year 1 = PRICING-D2 pilot tier (below the menu floor)
• Option C structural shape (flat + per-integration) is the closest match to the multi-phase growth model Andy validated

Negotiation tactic: Andy explicitly said leadership would not negotiate hard on the pilot number — "they would leave [pricing] up to me anyway. And I'm not going to do that." Pricing flows through Andy, not directly through Brian/Ed/CFO conversations.

Most pressing competitive intel#

Canals' public pitch line ("captures the institutional knowledge of your best talent before they retire") is essentially ours. Need to know if Brehob has been pitched — added to Andy meeting #2 prep doc as a question. If yes, our deck differentiation against Canals needs to be crisp on the corpus-depth dimension. If no (likely — small-mid market industrial distributors typically don't run vendor RFPs), we have first-mover advantage on the institutional-knowledge framing for Brehob specifically.

What stays out of this section (lives in competitive-landscape.md)#

• Per-category vendor tables and deep dives (5 categories × 6-12 vendors each)
• Direct competitor teardown: Canals (side-by-side comparison)
• Adjacent tools / integration targets — industrial distributor ERPs. Updated 2026-05-05: Brehob is on Microsoft Dynamics 365 Business Central (confirmed in Andy meeting #2), not Epicor Prophet 21 as previously hypothesized. The P21 first-mover MCP-server thesis remains valid for the broader market and Hannah Labs Customer #2 prospecting, but is N/A for Brehob specifically.
• Monitoring + revisit triggers
• Source links

Pricing Strategy#

Updated 2026-04-30 after value-model deep-dive (sub-agent margin analysis on 42 Brehob pricing sheets) and John call. Working numbers supersede earlier $5K/mo placeholder. See quoteai/pricing-model.xlsx for the live spreadsheet.

Value Framework — Five Drivers#

Updated 2026-05-04 with expanded driver framework from chat brainstorming session and the 3-person sales support team finding from John. The "five drivers" headline name is preserved because it reflects the deck's leadership-friendly simplification (Slides 8-12); the underlying analysis below tracks 11 drivers across four mechanisms. Stack-summed scenarios remain consistent with the value model: Conservative ~$2M / Mid ~$4M / Bullish ~$6M.

QuoteAI delivers value through four mechanisms — speed, accuracy, consistency, and (NEW per platform reframe 2026-05-04) searchability/queryability — converting into 11 distinct outcomes. Each outcome has a different certainty profile and beneficiary, which informs how it gets priced (fixed retainer covers high-certainty drivers; variable percentage shares the upside on uncertain drivers).

Mechanisms (HOW value is delivered):

• Speed — drafts in 90-110s vs the current 48hr-5day turnaround
• Accuracy — catalog matching + structured pricing + verbatim phrasing reduces errors
• Consistency — same template, same phrasing, same structure across all reps and quotes
• Searchability / queryability — vector + structured retrieval over the corpus, exposed via chat (the platform thesis; chat ships pre-trip per leadership-deck Slide 6)

Outcomes (WHAT Brehob gets — full 11-driver table):

Important: levers 6 (democratization) and 8 (close rate) overlap. Democratization manifests partly through close-rate uplift across the 17 non-John reps. For Brehob mid-scenario sizing, pick ONE lens — sized via #8 ($1.7M at +2pp close rate) is the safer/measurable lens. The democratization frame is the bigger number ($5M at 25% gap close) but harder to attribute. Don't add both — that overstates value.

Stacked mid-scenario annual value (lens #8):

Conservative case ~$2M (just #1 + #5 + #9 — the most defensible drivers). Bullish ~$6M+ (lens #6 democratization + close rate + completeness compounding). These ranges remain consistent with pricing-model.xlsx (Conservative ~$1.89M / Mid ~$3.79M / Bullish ~$6.19M).

Mapping to the deck "Five Drivers" simplification (Slides 8-12):

Both views produce consistent stack totals matching the value-model spreadsheet. Keep the deck simple; keep this doc nuanced.

Driver-by-driver value defense (for negotiation pushback):

• Knowledge transfer/insurance: John's $2-5M annual book attrites without active retention. QuoteAI captures his quote DNA so any salesperson sounds like John on his accounts. Insurance against a definite cliff. Plus: the same knowledge base, queryable via chat, lets anyone in the company access institutional memory beyond just quoting (lever 10).
• Throughput: 18 reps × 10 quotes/mo = 2,160/yr today. Even +20% throughput at 13% close × $40k avg deal × 35% margin = $470K/yr GP.
• Win-rate uplift: John identified speed as the #1 lost-deal reason. Brehob's 48hr target vs. QuoteAI's <4hr. 1% close-rate lift = $302K/yr GP. Knowledge democratization (lever 6) drives an additional but overlapping uplift on this same dimension — pick one lens for sizing.
• Cost savings: ~2.5 hrs/quote × 2,160 × $60 loaded × 75% realization = $243K/yr direct savings. Concentrated in the 3-person sales support team that builds and reviews every quote — see next section for the cost-replacement detail.
• Quote completeness: Per John, "shocked how many people don't bother with installation." Installation is 33-50% of project value; PM/parts carry 40% margin. Auto-attach captures revenue most reps leave on the table today.
• (NEW) Institutional queryability (chat): anyone in the company asks the corpus questions in natural language. Same knowledge layer, new use case, no new integrations required. Hard to size in dollars; high strategic value as a platform-thesis proof.
• (NEW, deferred) Strategic intelligence: Year 2+ extensions (CRM, inventory, territory) layer additional data sources onto the same knowledge platform. Sized when those integrations land.

The 3-Person Sales Support Team — Cost Replacement Detail#

Source: John call 2026-05-04. Brehob runs a 3-person sales support team — one per region (North/Central/South) — that builds and reviews every quote across the company. Per John: the team is hard to work with, has consistent absenteeism, and produces inconsistent output (turnaround swings based on which support person is working). This finding materially defends the "labor savings" lever (row 1 of the value table above).

Cost calculation:

Sizing scenarios for the labor-savings driver (row 1):

The deck's Slide 12 number ($240K) lands at the upper end of Mid — defensible because:

• The deck math (5,400 hrs × $60/hr × 75% realization = $243K) gives the same number the FTE-based math gives at ~2.5 FTE × ~$95K loaded
• Both math paths converge on $240K as a defensible mid-scenario floor
• The 3-person team mechanism makes this floor concrete and tied to a known operational pain point that leadership already feels

Framing for leadership conversation:

Don't lead with "fire 3 people." That signals badly AND it's not quite accurate — some of the team's time is on customer support / coordination that QuoteAI doesn't replace. Lead with the role evolution:

"Your sales support team currently spends most of their time on quote building. With QuoteAI, that workload shrinks dramatically — they can refocus on higher-value account/customer work, or the team can naturally consolidate as roles evolve. Either way, the cost of the current quote-building bottleneck is recovered."

Pricing implication:

The labor-savings driver ALONE defends the conservative-scenario value of ~$2M when combined with completeness ($800K) + insurance ($250K) + onboarding ($150K) + accuracy ($60K) — even before any close-rate or throughput uplift materializes. The 3-person team mechanism makes this floor concrete enough to anchor:

• Option B's $12-13K/mo flat retainer (captures ~6% of the conservative scenario's value)
• Option C's $5-7K/mo + growth fee structure (lower base recovered via growth)

For the leadership conversation, this is the CFO-resonant single-line case when other drivers are challenged:

"Even if we discount throughput and win-rate uplift entirely as speculative, the recovered labor cost from the support team alone — combined with quote completeness and knowledge insurance — justifies the platform fee. Everything else is upside."

Two Pricing Options#

Updated 2026-05-05 post-Andy meeting #2 — pilot pricing now lives in decisions.md § PRICING-D2. The four-option menu below (A/B/C/D) describes structural options for negotiation, but for Brehob first-customer specifically, Andy's read pointed to a "price to get in" tier intentionally below the menu.

Pilot pricing (Brehob, Year 1) — supersedes the menu:

Andy's framing: "I would price to get in. Basically get it in there, get it, you know, let people get it in, break it, figure it out, learn from it. Then you're going to have a more better refined product." Multi-phase pricing growth via scope expansion (each new phase = new pricing conversation) brings Year 2+ pricing into the comp range over time.

Cost-validation gate: before locking $3,500/mo, run an honest AWS cost analysis (Bedrock, RDS+pgvector, S3, ECS/Fargate, monitoring, embeddings, chat usage). If true cost exceeds $500/mo assumption, the multiple still applies — price floor moves up. Target: complete cost analysis BEFORE May 12 pitch so numbers are defensible if Brian asks.

Negotiation expectations: Andy will not let leadership negotiate hard — "they would leave that up to me anyway. And I'm not going to do that." Pricing flows through Andy.

Below: structural option menu (still relevant for scaled customers / Customer #2+ / Brehob Year 2+ scope expansion). Four pricing options, each tied to a different procurement persona and value-prop emphasis. Different procurement personas respond to different structures, and giving leadership a menu beats handing them a single take-it-or-leave-it.

Option A — Pure Flat (the safe option for variable-averse CFOs)

• Simplest to budget, no upside fights, no attribution disputes
• Captures the conservative-case value at ~15% of $2M
• Loses upside on bullish scenarios entirely
• Best for: procurement-led purchasing where predictability > alignment

Option B — Hybrid Capped

• Aligned incentives via variable component
• CFO comfort via cap
• Variable rate (6%) anchors to Brehob's own salesperson commission rates ("half a salesperson's commission for the assist")
• Sacrifices tail upside above $25M+ attributed sales
• Best for: CFO who wants alignment but bounded total spend

Year-1 expected totals at different attribution levels (35% margin assumed):

Option C — Hybrid Uncapped on Growth

• Lower base than Option B; upside captured via growth fee + per-integration economics
• "GP growth above baseline" is harder to dispute than "QuoteAI-attributed" (no annual fight about counterfactual; baseline locked in MSA)
• Per-integration economics scale naturally with the platform thesis (chat, inventory, CRM, etc.)
• Year-1 ARR scales smaller initially (~$80-150K) but compounds with each integration
• Best for: growth-oriented leadership who want vendor incentivized on outcomes; multi-year vision

Note 2026-05-05: Option C's flat retainer component ($5-7K/mo) is the closest structural match to PRICING-D2 pilot pricing ($3,500/mo). Andy's read pushed the floor lower than even Option C's low end — "price to get in" trumps comp-anchoring for first-customer dynamics. Option C remains the right shape for post-pilot Brehob Year 2+ as integrations layer in.

Option D — Outcome Packages (bundled framing for outcome-buyers)

Package integrations into named outcomes with bundled pricing:

• Sells outcomes, not features — leadership decks love this
• Per-integration economics hidden inside the package
• Easier to renegotiate at package boundary than per-integration
• Internally still structured like Option C; the package is a marketing wrapper for procurement
• Best for: sales conversations where "what does this thing DO for us" beats "how much does this cost"

Recommendations (revised 2026-05-05):

Walk-away floor (Brehob pilot): the $3,500/mo + $5-10K is already at-or-below comp range; deeper discounting would put the engagement underwater. If true AWS cost analysis pushes the cost-floor higher, the multiple still applies (×5-7) — the price moves up, not down.

Anchor positions (post-Brehob, future customer prospecting):

• Opening: Option C at $6K/mo + 2.5% YoY GP growth. Show menu (A-D); recommend C.
• Step-down: Option B at $12-13K/mo + 6% capped at $40K/mo.
• Last resort: Option A at $25K/mo flat (concedes alignment to win the deal).
• Walk-away floor (future customers): $5K/mo + 4% GP. Any worse, decline.

AWS Cost Validation (PRICING-D2 lock 2026-05-05)#

Cost-validation gate for PRICING-D2 ($3,500/mo + $12,500 setup). Closed 2026-05-05 with full breakdown below. Source for the locked numbers in decisions.md § PRICING-D2.

Workload assumptions (Brehob Year 1):

• ~40 active users (18 air reps + ~12 CSTs + ~10 leadership)
• Quote drafting: ~120 quotes/mo (air + motors only per Phase 1; ~60-70% of company total ~180)
• Per quote: ~25K input + 12K output Sonnet tokens
• Chat: 40 users × 5-10 questions/day × 22 working days = ~6,000 questions/mo
• Per chat: ~5-10K context input + 500-2K output (Sonnet synthesis; vector retrieval pure-pgvector, no LLM)
• Corpus: ~5K initial Phase 1 docs (John's archive + Andy's 3 boxes), grows ~120 quotes/mo

Monthly operating cost (with Y1 optimization roadmap)#

Multiple analysis at $3,500/mo price#

The trade: ×4.5 at mid is below the ×5 ideal but defensible per Andy's "price to get in" framing — pilot pricing trades margin for foothold. The high-scenario thin margin is the real risk; Haiku-retrieval validation (Y1.5) is the lever that unlocks headroom for that case.

Y1 optimization roadmap (must-ship)#

Y1.5 optimization (validate before deploy)#

Haiku 4.5 for retrieval/reranking (deferred from Y1 must-ship per Dan 2026-05-05 — quality unvalidated):

• The chat spike runs Sonnet end-to-end; quality of the demo moments is partly Sonnet's reasoning, not measured separately
• Two unvalidated sub-patterns: (a) replace Sonnet-as-agent with Haiku-as-agent; (b) add Haiku rerank between vector search and Sonnet synthesis
• Risk: Haiku might miss subtle relevance matches on technical/named-entity queries (compressor models, NFPA refs, customer names)

Validation plan (~half a day post-trip):

1. Build query panel — 15-20 representative questions:
   • 5 demo-tested (hospital, food-grade, centrifugal — already proven)
   • 5 leadership-likely (KPI/dashboard-style, throughput inquiries)
   • 5 sales-ops style (vendor-specific, model-variant, warranty)
   • 5 edge cases (long context, multi-hop, deliberately ambiguous)
2. Run two pipelines per query: A) Sonnet end-to-end (current spike) vs B) Haiku-rerank + Sonnet-synthesize
3. Score on source recall, synthesis quality, latency
4. Decision rule: if B matches A within tolerance on ≥18/20, deploy. Otherwise hold and revisit with prompt-engineering on the rerank stage.

If validated, expected Y1.5 cost savings: chat $540 → $400-450/mo. Mid total ~$631-680/mo. Multiple at $3,500/mo improves to ×5.1-5.5 (back into the comfort range). High scenario also improves materially.

Per-doc ingestion cost formula (setup-fee math)#

Hard AWS/API cost per doc (negligible):

Dan-time per doc (the real cost driver):

• Avg ~30s supervision/spot-checking on clean docs
• Edge cases (10-20%): ~5min each (parser failures, manual classification, weird templates)
• Net: ~1min average per doc supervision
• At $150-200/hr internal rate → $2.50-3.30/doc

Customer-facing tier (with margin recovery): ~$1.50/doc charged to Brehob — under-recovers Dan's actual time but anchors at a Brehob-digestible level. The under-recovery amortizes through the monthly fee stream.

Setup fee tiers#

Note on auth absorption: the ~30-55 hours of Cognito + Entra federation work (per INFRA-D3) is baked into the $5K base setup, accepted as under-recovered first-customer build cost. Customer #2+ inherits the productized auth pattern — their auth setup labor is a fraction of the first-customer cost. This justifies the $5K base feeling slightly higher than the AWS-hard-cost line ($1-2K) would alone suggest.

What's NOT included in setup (out-of-scope, separately scoped):

• Business Central API integration (Phase 2/3)
• Dynamics 365 CRM integration (Phase 2/3)
• Datasheet auto-attach (Phase 2 — asset-curation workflow)
• KPI dashboard expansion beyond v1 (Phase 1 ships a minimum viable dashboard pre-trip)

Monthly cost-tracking deliverables#

Per the MSA cost-transparency principle (introduced in Usage / API Consumption Considerations), Brehob receives a quarterly statement covering:

• Queries/mo (chat + drafting)
• API spend by component (Bedrock chat, drafting, embeddings)
• AWS infrastructure spend (RDS, ECS, NAT, etc.)
• Allowance utilization (if/when allowance tier introduced post-Y1)

Transparency builds trust and makes the Year-2 conversation about expanded usage smoother — no surprises.

Risk callouts#

• Heavy chat usage from leadership demos / curiosity traffic could push toward high scenario in the first 3-6 months. Mitigation: daily budget alarm + sliding window.
• Bedrock model lag (per INFRA-D2) means we eat any temporary cost premium until parity ships. Default position: absorb for Brehob pilot.
• RDS scale-up (multi-AZ, larger instance) at customer growth points adds ~$150-250/mo. Likely Y2 conversation, not Y1.
• Cognito free-tier breakpoint at 50 federated MAUs — Brehob is at 40; user-count growth past ~50 triggers $0.015/MAU billing (still negligible: $5-10/mo at 100-150 users).
• Haiku-retrieval validation failing — chat-cost margin stays at ×4.5 mid / ×2.3 high without the optimization. Mitigation: fall back to Sonnet end-to-end and accept the thinner margin in high-usage months.

Per-Integration Economics#

Year 2+ extensions need their own pricing logic since they layer onto the platform foundation differently. Captured here as a sub-framework that Options C and D depend on.

For Year 2+ extensions (inventory, CRM, equipment library, customer portal, territory intel) the pricing layers should map to integration value type:

Key principle: cost-savers (inventory, equipment catalog) get fixed pricing; revenue-generators (CRM, territory intel) tilt heavily toward performance pricing. Don't apply a single pricing model uniformly — match the model to the value type.

Build-fee economics for Hannah Labs as multi-customer co.:

The first customer pays a higher build fee to fund the engineering work. Once integrations are productized via Forge (per projects/forge/design.md), subsequent customers pay a lower or near-zero build fee (it's already built) but the same monthly + performance structure. This is how the per-integration model future-proofs for Hannah Labs growth:

This pattern is also how the AWS upfront cost (per INFRA-D2) gets passed through cleanly: build fees compensate engineering labor; AWS fixed costs amortize across the customer base over time.

MSA scope for integrations: each Year 2+ integration is a separately-scoped project per the MSA Scope subsection below. Brehob's foundation MSA covers quoting + chat + KPI dashboard + datasheet auto-attach (Slide 15 in-scope list). Year 2 integrations are project-based add-ons evaluated and priced when their time comes — not pre-bought via the foundation retainer.

Usage / API Consumption Considerations#

Added 2026-05-04 with the chat extension shipping pre-trip and AWS-native deploy direction (INFRA-D2). Chat exposes materially higher API usage than quote drafting alone, which means API costs need explicit accounting in the pricing model.

Usage profile shift with chat:

• Quote drafting (existing): ~1 generation per quote, batched, ~20-30K input + 10-15K output tokens. Predictable per-quote cost.
• Chat (new pre-trip): many small calls per session — retrieval + reranking + final response, repeated per question. Highly variable per-session cost.

Order of magnitude estimate (Brehob-scale):

• 18 reps + leadership using chat at 5-10 questions/day each
• Token cost via Bedrock (Sonnet for synthesis, Haiku for retrieval/rerank): ~$0.05-0.30 per question depending on context size
• Per-month spend: ~$150-1,500/mo

That's manageable in absolute terms but high enough that it can't disappear into platform-fee margin without explicit accounting — especially when AWS upfront costs (per INFRA-D2) are also being passed through.

Three approaches for handling usage in pricing:

Recommended for Brehob (first customer): Approach 1 (allowance + overage) baked into whichever pricing option (A/B/C/D) is selected.

Reasons:

• Predictability matters more than perfect alignment for first-customer trust-building
• The platform fee already does some passthrough of AWS infra cost
• Brehob has known user count (18 reps + leadership) so allowance can be calibrated reasonably accurately upfront

Allowance sizing for Brehob: ~3-5K queries/mo across 18 reps + leadership = ~$300-500/mo of API spend covered by allowance. Overage tier above at $0.10/query (covers cost + small margin).

At Hannah Labs scale (multiple customers): revisit toward Approach 2 or 3 once usage variance across customers becomes apparent. Approach 1's flat allowance becomes unfair when one customer uses 10x another. Likely a Y2 conversation, not Y1.

Open question to lock with Andy/CFO:

When Bedrock model lag (per INFRA-D2 trade-offs) creates a temporary cost premium (e.g., direct API has a cheaper Haiku tier that Bedrock hasn't matched yet), do we eat the premium or pass through? Default position: eat it for the Brehob pilot since the cost differential is small and worth not introducing variability. Revisit at Year 2 review.

Cost transparency principle for the MSA:

Quarterly statement to Brehob: queries/mo, API spend, allowance utilization, overage if any. Transparency builds trust and makes the Year 2 conversation about expanded usage smoother — no surprises.

Negotiation Playbook#

Bundle the price; decompose the value defense.

The five drivers above are NOT line items in the offer. They're ammunition for defense. Why:

• Putting them as separate line items invites cherry-picking ("we don't believe in driver 1, cut $X")
• Each driver is harder to defend individually than the bundle
• Attribution disputes per-component create contractual complexity

Tactic:

• Quote ONE bundled offer (Option A or B)
• HAVE the per-component value breakdown ready
• When pushed back on total: "Which of these five drivers do you not see real value in?"
• Burden flips: they argue against specific value, not "your number is too high"

Anchor positions:

• Opening: Option A at $12-13k/mo flat + 6% GP variable
• Walk-away floor: $5k/mo flat + 4% GP variable. Below this, decline and revisit later.
• Year 2/3 escalators: baked into MSA — flat $14k → $16k → $18k over 3 years; variable % unchanged (escalating that feels like a rug pull).

MSA Scope — In and Out#

The MSA covers ONE thing: quoting. Everything else is project-based add-on.

Why scope-narrow matters: scope creep kills MSAs. Integrations have variable cost (Brehob's ERP could be a 6-week build or 6-month nightmare — can't estimate without scoping). Year 2 conversation becomes "Inventory integration would be a separate $80k project; we estimate it lifts QuoteAI value by $X/yr — worth it?" Real business expansion, not feature creep.

MSA language sketch:

"Scope: development, hosting, and ongoing engineering of the QuoteAI quote-generation tool, including all features described in [Schedule A]. Integrations with external systems (CRM, ERP, inventory, accounting) are not included and will be separately scoped on a project basis. Custom feature requests outside the quarterly release cadence will be evaluated and quoted separately."

Pending Inputs#

Numbers to tighten before final pricing lock-in:

Resolved 2026-05-04 (this update):

• Typical hands-on-keyboard time per quote (rep + support) → John confirmed 3-person sales support team builds + reviews every quote; team is the labor-savings concentration point. Cost replacement detail captured in the new sub-section below.
• Customer loyalty split (loyal-to-John vs loyal-to-Brehob) → still valuable to tighten knowledge-insurance lever (#5) but no longer blocking the pricing menu structure.

Still pending:

• From Matt (Brehob central-region ops): total monthly quote volume team-wide, win rate by region, avg deal size by region. Initially due 2026-05-01; chase before Indianapolis trip if not already received.
• From Andy: Brehob compressor-division revenue split (% of $50M total — current working assumption is ~$30M, 60%), strategic priorities for next 12-18 months (drives which value driver leads the deck), CFO budget appetite signal, in-room composition (CFO heavy vs. ops heavy → drives pricing-option emphasis).
• From John (followup): customer loyalty split (loyal-to-John vs loyal-to-Brehob — directly tightens knowledge-transfer driver); 3-person support team specific roles (which one is per-region anchor; whether headcount has fluctuated); any prior outsourcing experiments with quote-building.
• 3-year trailing GP baseline for Option C: if the recommended pricing model (Option C — uncapped on growth) is the lead anchor, lock the baseline calculation before any contract conversation. Recommend trailing 3-year average GP from the air division specifically (not whole-company) so growth math doesn't get muddied by other-division dynamics.

When these land, refresh pricing-model.xlsx Inputs tab; this section's anchor numbers may shift modestly but the structural option menu (A/B/C/D) stays.

Revisit Triggers#

• 12-month review with measured-attribution data. Renegotiate based on actuals, not pre-trial estimates.
• If QuoteAI-attributed sales exceed $20M annually: trigger early renegotiation. Both parties win — variable cap may need adjustment, and Brehob may want to expand scope.
• If adoption stalls (attributed sales < $1M after 6 months): trigger early review to reset scope, price, or both. Worst case: terminate cleanly per § Contract Shape § 3.
• Year 2 step-up automatic unless either party raises a concern at the 12-month review.

Working References#

• quoteai/pricing-model.xlsx — value model spreadsheet (5 tabs: Inputs, Value Model, Notes & Sources, Margin Reference, Sensitivity Analysis)
• quoteai/pricing-summary.md — sub-agent margin analysis on 42 Brehob pricing spreadsheets (rotary 44%, dryer 45%, vacuum 31%, parts 41% list-margin)
• quoteai/pricing-analysis.csv — 209 line items with multipliers, browseable
• quoteai/decisions.md § PITCH-D2 — engagement framing (software with dedicated engineering support, not consulting)
• quoteai/decisions.md § E4-D2 — commission category structure (informs the 6% GP anchor)

The Extensible Pipeline#

Product framed as a 4-stage pipeline with pluggable stages:

Collect → Draft → Review → Done

What the MVP delivers today#

All four stages are live end-to-end. This isn't a mockup or a slide deck — it's working software.

• Collect — existing form (E3) + spreadsheet upload (E4, in progress)
• Draft — Claude Agent SDK fires MCP tools (search_line_items, search_past_quotes, search_equipment) to assemble the draft from 22 years of Brehob's quote history
• Review — three-gate annotation system (amber caveat / red review / blue price) + inline hover popovers + chip navigator + pricing MetaRail (E4) surfacing commission take-home
• Done — rendered markdown quote with Brehob letterhead, copy-to-clipboard

Candidate future extensions#

Each becomes a possible deliverable for a future month of retainer work. Quote Log is the lead — John specifically mentioned it, and it directly answers Andy's "how does this impact the business?" question.

• Quote Log + Sales Leadership Dashboard — unified sales-activity surface built on top of the quote data: which salespeople made what quotes for which customers, total quoted amounts, pipeline status, commission distribution across the team, historical quote search. Consolidates what were earlier listed as separate items (quote analytics, commission visibility, historical viewer) into one coherent feature. Directly visualizes business impact; highest-leverage near-term candidate for the Andy pitch + leadership conversation.
• Complete proposal packaging — quote + diagrams + supporting materials (catalog sheets, warranty info, installation diagrams) bundled as one deliverable
• Inventory integration — real-time stock checks, availability constraints, alternative-product suggestions when items are out of stock
• Preventive maintenance quotes — separate quote type for service contracts / recurring maintenance (different template, recurring-billing data model)
• Customer-facing quote portal — customer opens a link, views the quote, asks clarifying questions in-thread, e-signs. Replaces email back-and-forth.
• CRM integration — push closed quotes to Salesforce / Hubspot / etc.
• AI-assisted objection handling — when a customer pushes back on price or timing, tool surfaces responses based on past-quote data where similar objections were handled successfully
• Multi-region / multi-territory support — if Brehob expands beyond current territories

The shape of the roadmap#

Brehob chooses the order based on their priorities, not a vendor-dictated roadmap. Each future month is scoped with Andy (or successor relationship owner) at the start of the month. That's the "enhance don't change" principle at the engagement level — Brehob owns the priority stack.

Contract Shape#

The five knobs Andy (and eventually leadership) will ask about. Not all need to be surfaced in the Andy meeting itself (PITCH-D1 — meeting is champion-enablement, not deal-close), but they need pre-decided answers so any can be addressed confidently.

1. What does $5K buy in a given month?#

• (a) Unlimited — "Dan handles whatever Brehob asks for"
• (b) Hour-capped — "Up to 40 hrs/mo; extra rolls to next month"
• (c) Deliverable-based — "Monthly scope conversation; Dan commits to 3-5 deliverables"

Working answer: (c) deliverable-based. Aligns with "embedded engineer" framing; each month kicks off with a 30-min scope conversation. If Brehob asks for more mid-month, it bumps to next. Hour caps feel transactional (contractor-y); unlimited is a burnout trap.

2. Pricing structure#

Updated 2026-04-30 — see ## Pricing Strategy above for full analysis.

• (a) Flat $5K/mo — original working answer; superseded
• (b) Per-quote — rejected (measurement overhead, "is Dan gaming this?" anxiety)
• (c) Hybrid: flat + % of generated value, capped — NOW RECOMMENDED

Working answer: Option A from ## Pricing Strategy — hybrid $12-13k/mo flat + 6% of QuoteAI-attributed gross profit, capped at $40k/mo total. Aligns with Brehob's own salesperson commission rates (anchor: "half a salesperson's commission for the assist"). CFO can budget the flat; variable only fires when QuoteAI is actually generating sales.

Alternative for variable-averse CFOs: Option B pure flat at $25-30k/mo. Present as good/better menu, not à la carte.

3. Termination#

• 30-day notice on either side (default)
• Brehob can terminate for cause (non-delivery) without notice
• Dan can terminate for non-payment after 14 days

4. Renewal after trial#

• Month-to-month; no auto-escalation
• $5K stays until explicitly renegotiated
• Quarterly "scope + price check-in" with Andy (or whoever owns the relationship)

5. Priority tier as Dan's portfolio grows#

• During trial + first year: Brehob is priority-1 (< 24hr urgent, < 1wk normal)
• If Dan signs other clients later: Brehob stays priority-1 unless mutually agreed otherwise

These 5 knobs become a one-page MSA-light for leadership approval after the Andy meeting.

Risks & Mitigations#

IP portability#

• Concern: without careful contracting, Brehob could claim exclusive rights, blocking resale to other distributors.
• Position: Dan retains code + domain learnings; Brehob gets perpetual, non-exclusive use on their own infrastructure. Language TBD.

Monthly deliverable legibility#

• Concern: "incremental value" is squishy; Andy needs to feel each $5K is earned.
• Mitigation: monthly scorecard — quotes generated, time saved, errors caught, commission $ surfaced. Establishes review cadence from month 1.

Dependency risk ("what if Dan disappears?")#

• Mitigation: documented handoff plan, code in Brehob-visible private repo, pre-agreed buy-out price for ownership transfer.

Scope creep from Brehob#

• Concern: Brehob requests accumulate faster than Dan can deliver. Without process, the retainer becomes a rolling pile of "quick asks" and Dan burns out.
• Mitigation: Monthly scope conversation with Andy (ties to deliverable-based scope — see Contract Shape § 1). New mid-month asks queue for next month's scope instead of interrupting.

Dan's capacity as portfolio grows#

• Concern: Dan takes on a second client; Brehob's response times degrade without explicit priority hierarchy.
• Mitigation: Priority-1 tier baked into contract (< 24hr urgent, < 1wk normal) during trial + first year. Concurrent client engagements don't drop Brehob from priority-1 unless Brehob mutually agrees.

Data security#

• Concern: Customer PII (names, phones, addresses), pricing, commissions all flow through the system. A breach is both legal liability and Brehob-relationship-ending.
• Current state: app is local-only on Dan's dev machine. Not a real risk for the demo, but becomes one the moment the tool runs on anything persistent at Brehob.
• Mitigation: Before hosted deploy, commit to: encryption at rest, no PII in logs, Brehob-controlled backup schedule, written incident-response SOP. Blocker for "post-MVP hosted deploy" scope.

Time horizons: IP portability matters at contract signing. Scope creep + capacity are ongoing management disciplines. Data security matters at hosted-deploy time. Different risks need answers at different moments.

Pricing & Success Metrics#

Pricing#

Superseded 2026-04-30 — see ## Pricing Strategy above for current numbers.

The $5K/mo placeholder below was the working number from John's 2026-04-22 conversation. After the value-model deep-dive (sub-agent margin analysis + John follow-up call 2026-04-30), the recommended pricing is materially higher and structurally different (hybrid flat + GP variable, not pure flat). Kept here for the decision trail.

Why the change: the $60K/yr value claim turned out to be a 30x undercount. Real value drivers (knowledge transfer + throughput + win rate + cost savings + quote completeness) compound multiplicatively at industrial-distribution scale. See ## Pricing Strategy for full analysis.

Success Metrics (post-Andy, pre-trial)#

Once leadership approves the trial, refine these with Andy + sales leadership before month 1:

Quantitative (measured from the app)

• ≥ N quotes generated per month (anchor: John's "good year = 150/salesperson/yr"; target uplift)
• ≥ X% of generated quotes close to revenue (tagged manually in the tool)
• Average quote turnaround < Y minutes (baseline today: ~48 hours)
• Per-salesperson quote throughput delta month-over-month

Qualitative (surveyed monthly)

• Salesperson NPS > +30 (or delta from baseline)
• Leadership "would you renew at current price" — yes / no / maybe-at-lower-price

Renewal gate at month 3

• Quant hit + qual positive → auto-renew at $5K/mo
• Quant miss + qual positive → renew at negotiated rate ($3-4K while scope re-aligns)
• Qual negative → terminate, Brehob keeps code via buy-out clause

Open: realistic targets for N / X / Y come from a 30-day baseline after onboarding, not pre-trial guesses. Commit to measuring baseline + setting targets with leadership sign-off.

Stakeholder Framework#

Brehob's decision-making isn't a single-person gate.

Confirmed#

• John — champion, brings Dan to Andy, first internal advocate
• Andy — first decision-maker; brings QuoteAI to leadership for approval (not the final gate)

Likely exists at Brehob (confirm via John)#

• Leadership team (C-suite or directors) — actual approval gate for the retainer
• IT / security lead — will want to review tech stack, data handling, deployment architecture. Involvement triggers when moving to hosted deploy (post-demo).
• Sales leadership (VP Sales?) — manages the salespeople. Buy-in = adoption; resistance = dead on arrival regardless of leadership approval.
• Finance / procurement — $5K/mo is likely a director-approval line item. Invoice structure needs to fit their AP systems.
• Legal — for MSA / NDA. Brehob may have a standard vendor template.

Questions for John before the Andy meeting#

• Who else at Brehob would need to approve / review this beyond Andy?
• Any known objections or concerns from the sales team?
• What's the standard vendor onboarding process (NDA, W-9, insurance, etc.)?
• Does Brehob have a preferred contract / payment terms template?
• What's the cultural read on AI / software tools at Brehob?
• Per-quote commission estimate — to model the throughput-uplift ROI story

Open Questions#

• Right opening price anchor — $5K, $7.5K, higher? Resolved 2026-04-30 — see ## Pricing Strategy. Opening: $12-13k/mo flat + 6% GP variable. Floor: $5k/mo flat.
• Exclusivity windows — any value, or avoid?
• Trademark / domain check on "QuoteAI" product name (several tiny GitHub repos already use it).
• What does Brehob consider a "win" after 12 months? (Becomes the refined Success Metrics target — see ## Pricing & Success Metrics.)
• NEW (pending Matt 2026-05-01): team-wide quote volume + win rate + avg deal size by region. Tightens value model inputs for final pricing lock-in.
• NEW (pending Andy): Brehob compressor-division revenue split — anchors % of revenue ceiling test for pricing.
• NEW: customer-loyalty split for John's accounts (loyal-to-John vs loyal-to-Brehob) — directly tightens knowledge-transfer driver in the value model.

Update this doc after every stakeholder conversation. Meeting-specific tactics live in andy-meeting.md or successor meeting docs.