Complex quoting, made conversational.

Sellers and partners describe the deal in Slack, Teams, or the CRM. Across returns a configuration and a draft quote that follow your rules.

Request a demo

deal-desk

  1. Jordan Lee

    Renew Northwind for a year, add 500 Analytics seats, via Summit Partners.

  2. Across

    Found the contract (ends Mar 31) and the partner (Summit, Gold).

    Start the new seats now, prorated, or at renewal?

  3. Draft quote · Northwind

    • Analytics · 500 seats Prorated 5/12 $26,250
    • Renewal · Platform + Analytics 12 months $183,000
    • Gold partner discount 12% −$25,110

    Total $184,140

    Standard partner discount. No approval needed.

    Submitted to your CPQ as a draft.

    1. R-20 Add-ons co-term, prorated by month Deal desk walkthrough
    2. R-12 Gold partner, fulfillment deal: 12% off Partner guide, p. 4
    3. R-13 Above partner rate needs director approval. Not triggered. Approval matrix, row 7

Where quoting breaks today.

Today, a seller’s request passes through configuration, contracts, pricing, and approvals by hand. Each handoff can delay the quote or send the wrong price to the customer.

Seller in Slack

Renew Northwind and add 500 seats, through our partner.

  1. Configuration Seller guesses
  2. Contracts Manual math
  3. Pricing and partners Wrong discount
  4. Approvals Wrong approver
  5. Deal desk Queued, experts pulled in

Quote sent Days later

Rework

How Across builds this quote.

Across checks the product, contract, partner price, and approval rules, asks for missing details, then drafts the quote.

Action

  • Ask what’s missing
  • Draft the quote
  • Route approval

Reasoning

  • Check configuration
  • Co-term and prorate
  • Price and discount

Context

  • Contract and account
  • Partner tier
  • Price book

Sources

  • CRM
  • CPQ
  • Pricing policy

Built from how you quote.

Many quoting rules live with your deal desk, not in the CPQ. The Architect reads your pricing policies and past quotes, then asks your deal desk about the gaps.

  1. Pricing policies, approval rules, past quotes, and your deal desk’s answers.

    • Pricing policy
    • Approval matrix
    • Recorded walkthroughs
    • Past approved quotes
    • Architect asks:Do we prorate mid-contract seats by day or by month? Deal desk answers:By month, co-termed to the contract.
  2. The Architect drafts it in plain language. Your deal desk reviews it.

    • R-12 Partner guide, p. 4 Gold partner, fulfillment deal: 12% off
    • R-13 Approval matrix, row 7 Above partner rate: director approval
    • R-20 Deal desk walkthrough Add-ons co-term, prorated by month
  3. A versioned graph of small agents, each with one job.

    • Action Ask what’s missing · Draft the quote · Route approval
    • Reasoning Check configuration · Co-term and prorate · Price and discount
    • Context Contract and account · Partner tier · Price book
  4. The Operator catches rule changes. Your approved quotes re-run before a person signs off.

    • Renewal and upsell, Gold partner Matches
    • EMEA multi-year deal Matches
    • Over-limit discount Sent for approval
    • Gold partner discount: was 12%, now 14% Past quotes re-run ·Waiting for approval

How we’ll measure the pilot.

Targets we agree with you before the build, scored on quotes your deal desk already approved.

Drafts that build in your CPQ with the right SKUs, quantities, prices, and dates.

Average time from the seller’s first message to a confirmed draft.

The right discount and approval path for each scenario.

Apply real rule changes, then re-run every past scenario with no regressions.

Questions to askCheck the answer against your own cases

Questions to ask any quoting AI.

  1. Does it know what isn’t in the CPQ?

    Why it matters: Most quoting rules live with product managers, SEs, and the deal desk.

    Across: The graph is built from all of it, not just the catalog.

  2. Does it ask only for what’s missing?

    Why it matters: Sellers stop using tools that make them repeat what the CRM already knows.

    Across: It checks the CRM and contracts first, then asks one question.

  3. Why it matters: The same request can’t produce two different prices.

    Across: Pricing, discounts, and co-terms run as rules, in code.

    • 500 seats × $126 × 5/12=$26,250
  4. What happens when a rule changes?

    Why it matters: Rules change every quarter. Each change can’t become an engineering ticket.

    Across: Describe it in plain language. It’s tested on past quotes and approved first.

  5. Can you trace every line to a rule?

    Why it matters: The deal desk and finance need to see why each number is there.

    Across: Every price, discount, and approval links to its rule and source.

Start with your hardest quotes.

A three-week pilot on your own data, scoped to one product family or region.

  1. Week 1
    • Catalog, pricing, and approval rules
    • 20–50 real scenarios with the approved quote
    • Two recorded walkthroughs
  2. Week 2
    • Writes the rulebook in plain language
    • Your deal desk reviews it
    • Compiles and tests the graph
  3. Week 3
    • Tested on held-out approved quotes
    • Your team tries it in a sandbox chat
    • Results against the agreed targets
  4. Your decision

    Connect your CPQ, CRM, and Slack or Teams for a first group of sellers.

What we need to start.

  • Product catalog SKUs, bundles, and descriptions
  • Pricing and discounts Price books, tiers, and partner rules
  • Contract rules Renewals, co-terms, and proration
  • Approval matrix Who approves what, at which thresholds
  • 20–50 real scenarios Requests with the approved quote
  • Two recorded walkthroughs With the deal desk and the CPQ admin
How it connectsPermissions, approvals, and your systems of record

Your systems stay the record.

Across sits between the conversation and your CPQ. It reads under each user’s permissions and writes only after a person confirms.

Across
  1. Permissions Reads only what each user can see
  2. Confirmation Writes a draft only after a yes
  3. Logged Every action is recorded
  1. Ask in Slack or Teams
  2. Read, under the user’s permissions CRM, contracts, price book
  3. Draft the quote back to the seller
  4. Write to the CPQ after a yes, and log it
  1. ITCleared
  2. InfoSecCleared
  3. RiskCleared
  4. SOC 2Type II
  • Cleared by IT, InfoSec, and risk teams at a Fortune 100 technology company.
  • Single sign-on through Okta, Entra ID, Google, or Duo.
  • Runs on the models you approve and does not train on your data.
across / new run start here

Let us help you get work done.