The customer is waiting. Sales wants speed. Engineering wants to double-check. Finance worries about margin. And someone is pasting numbers between spreadsheets like it's a hostage note.

I’ve sat in those calls. At 9:02 the rep says “Can I get a quick check on this valve package?” By 9:45, three seniors are debating a dependency that lives in a PDF from 2017. Nobody feels smart. Everyone feels stuck.

This isn’t a product problem. It’s a process that traps expertise in people’s heads and asks them to be human routers. That’s the real friction.

The Bottleneck Isn’t People. It’s Human-Gatekept Knowledge.

Most companies think the delay is caused by “slow engineering approvals” or “sales getting it wrong.” So they escalate more, add more reviews, and pull their most expensive people into more meetings.

What’s actually happening: you’ve forced your experts to run an internal helpdesk for rules and compatibility. You’re paying principal engineers to copy logic from their brains into chat threads. That is not value creation.

Stop promoting engineers to Human Middleware.

The bottleneck is tribal knowledge scattered across heads, PDFs, and spreadsheets. As long as the rules live there, quoting will feel like a negotiation with your own company.

CPQ is not about automation - it’s about correctness.

Speed is a side effect of getting the logic out of people and into a system you trust. You pave a road instead of relying on local guides. Anyone can travel it, even on day one.

Treat Expertise Like Code: Rules, Owners, Tests

If you want quotes in minutes without gambling on quality, treat product knowledge with the same discipline as software. One source of truth. Clear ownership. Versioning. Tests. The rest gets easier.

Here are the rules I use with teams that sell complex, configurable products.

  • Rule 1: Make logic the product. Your product knowledge is an asset. Model it as structured logic with explicit compatibility, exclusions, and required parts. Don’t bury it in slides. In practice: build a configuration core that sales and pricing consume. When talking about systems that only uses symbolic logic for configuration, rule quality beats UI every time.
  • Rule 2: Assign an owner, not a committee. Someone accountable for correctness and maintainability. Not a meeting. A name. In one med-tech program I ran, naming a logic owner cut quote escalations in half in the first quarter. Fewer cooks. Better soup.
  • Rule 3: Make the system explain itself. If a rep can’t see why a choice is invalid or a price moved, they won’t trust it. Show the rule, the dependency, the calculation. Explainability drives adoption. No black boxes.
  • Rule 4: Kill the anti-pattern. “Engineers as Quote Validators” is a smell. If experts touch every quote, you don’t have a system - you have a concierge service. Redirect that time to improving rules and tests so the next quote doesn’t need a person.
  • Rule 5: Correctness first, speed follows. Get your core combinations and price logic right, then iterate. Perfect pricing is a myth. Launch, learn, adjust. Quotes are perishable like fruit in a store. Every day on the shelf costs you.

Adoption is the only metric that matters.

One more point on shiny tools. AI won’t save you from missing rules. It depends on structure. With clear constraints, AI becomes a great assistant for explanations, documents, and comparisons. Without them, it produces fluent guesses. Don’t scale guesses.

What to do this quarter

  • Map the review loop. List every human touch between requirement intake and quote sent. Count the handoffs, not the days. Your delay lives in those touches. Remove the ones that check knowledge the system should hold.
  • Extract the top 20 rules. Take your most painful dependencies and encode them as structured logic, with tests. Run a pilot on one high-volume product family. Bring one rep, one engineer, one pricing lead into a weekly cadence. Ship learning, not slides.
  • Make a single source of truth real. Choose where the logic lives and who maintains it. Version it. Test it. Publish change notes to sales. If change still needs a project plan, the field will route around you.

I’ve done this with teams since 2000, from technical sales rooms at Tacton to global rollouts. The pattern is boring and it works: move expertise from people into rules, prove correctness, then scale. Sales stops stalling. Engineering gets their time back. Finance sleeps at night.

If you’re tempted to celebrate that senior engineer who “saved” another quote, check the system first. Saves are a signal your rules need work, not a medal for heroics.

Free your experts from gatekeeping and they’ll build you the road.

Stop asking your experts to review every quote. Start asking them to build the system that makes reviews obsolete.