Conga’s October 2026 release adds deal-level pricing controls to its CPQ product, so pricing policy applies to a whole quote instead of one line item at a time. The same release adds AI-assisted quoting and a server that lets Microsoft Copilot, Claude and other MCP-compatible assistants read quotes and approvals.
What Conga shipped
Conga set out the release in a post by Rob MacDonald, its associate vice president of product and portfolio marketing, published on October 14. The CPQ changes come in three parts. “New deal-level pricing controls help organizations govern pricing beyond the individual line item, ensuring larger and more complex opportunities remain aligned with corporate pricing policies,” the post says. Smart CPQ now connects to Rebate Management and Agreements, so sellers see the incentives that apply while they quote and can create or link rebates inside the quote-to-agreement process. A Quote Comparison view lets sellers and buyers look at several quote scenarios side by side.
The AI features sit next to those controls. Conga says its AiMe assistant now has more than 90 tools and 19 skills, with Quoting among them, and the new Conga MCP Server lets teams reach contracts, quotes, obligations and approvals in natural language from the assistants they already use. The post opens with the goal behind all of it: “Every organization is looking for ways to move faster without increasing risk.”
Advertisement
Why line-by-line discounting breaks down
Most CPQ approval logic was written for a person editing a quote. A rep changes the discount on one product, the system checks that line against a threshold, and an approval request goes to a manager. Conga’s post describes the alternative: “Rather than evaluating discounts one line at a time, organizations can apply pricing policies across the full quote, helping ensure large and complex deals remain aligned with margin objectives and approval processes.”
The difference matters most when software builds the quote. An assistant that adds an upsell product, applies a bundle price and attaches a rebate can produce a document in which every line passes its own check while the deal as a whole lands below the margin floor. A rule that reads the full quote catches that case, and a rule that reads one line at a time lets it through.
Conga also ties the change to work that happens after approval. Connecting pricing, incentives and agreements, the post says, gives “greater transparency into profitability before a deal is approved and less reliance on manual calculations, disconnected systems, and downstream reconciliation.”
ServiceNow’s CPQ head puts the rules first
Christopher Shutts, Group Vice President of CRM Sales & CPQ at ServiceNow, makes the same argument in his predictions for 2027, sent to SalesTech Edition by ServiceNow’s product public relations team. ServiceNow sells its own CPQ product, so the forecast comes from a vendor in the category Conga competes in.
“In 2027, generating a sales quote will become the easy part,” Shutts said in the comment. “As AI agents take on more responsibility across B2B sales, companies will be able to configure products, optimize pricing, add upsell products, and prepare proposals in seconds.”
His concern is what follows that speed. “The next evolution of CPQ will be about making sure every AI-generated quote is accurate, profitable, and aligned with the company’s commercial rules,” he said. “AI-powered selling will need to work within trusted configuration requirements, pricing policies, and approval processes, whether a deal originates with a salesperson, a partner, a customer, or an autonomous agent.”
He also names the cost of skipping that step: “Companies that rush to automate quoting without connecting these commercial rules will risk accelerating costly mistakes, from inaccurate pricing to promises that fulfillment teams can’t keep.”
The same order of work across vendors
AI quoting has reached several CPQ vendors this month. Threekit launched an AI-native CPQ that builds manufacturer quotes from a voice note or photo, and SAP showed AI-powered quoting inside its agent suite. Earlier, Salesforce and Infinitus each put a rules layer between the sales AI agent and the action it takes.
In our reading, Conga’s release and Shutts’s comment describe the same sequence: the pricing and approval rules come first, and the assistant that writes quotes runs on top of them. A product demo shows how fast an assistant produces a proposal. The release notes show where the controls sit, and the controls are what a pricing team works with every day after the purchase.
What it means for the sales leader
Start with an inventory of the rules your quotes depend on today: discount thresholds, bundle prices, rebate eligibility, product compatibility and who approves what. Mark which ones live in the CPQ system and which live in a spreadsheet, an email thread or a sales manager’s memory. Rules held outside the system never reach an AI assistant, because the assistant only sees what the system holds.
Next, test whether your approvals read the whole deal. Build a quote in which each line sits just inside its own discount limit and the total sits below your margin floor, then see whether the system stops it. A quote that passes is the case that grows as more quotes come from assistants, partners and customer self-service.
Then decide which channels may create a quote without a person reviewing it: reps using an assistant, partners, customers on a portal, or a buyer’s own agent. Shutts lists all four. Each needs the same price book and approval rules, and one person in revenue operations should own the record of which channel produced which quote.
Finally, ask every CPQ vendor in your next evaluation to run deal-level controls on a quote its own AI built, with a deliberate pricing error in it, and to show the approval that fires.
