Sales quotation automation without a full CPQ project.

Eighty-six percent of the 200 US manufacturers in Aleran's 2025 Built to Sell survey said they had lost deals because quoting was slow or manual. Only 22% of the same companies use CPQ software, the category built to fix exactly that. Read quickly, the two numbers say manufacturers are leaving money on the table out of stubbornness.

Read slowly, they say something more useful. Aleran sells CPQ, and the respondents who already know the pain have mostly looked at the standard cure and walked away. The reasons they walk away explain more about sales quotation automation than any feature list does, and they point to a smaller, faster fix that most companies can start on without replacing anything.

What is sales quotation automation?

Sales quotation automation is software doing the repeatable parts of producing a price offer: reading the request, looking up prices and terms, applying the discount rules, routing the quote for sign-off, sending it, and chasing it afterwards. People keep the parts that need judgment, such as a strategic discount or an unusual delivery promise. Everything rule-bound runs without anyone retyping it.

The phrase covers a wide spread of products. At one end is CPQ (configure, price, quote). CPQ is a full system that holds your product catalog, pricing logic and approval chain, usually inside a CRM such as Salesforce. At the other end are proposal tools that make a good-looking document and track when the buyer opens it. In between sits a lot of plain automation: an email arrives, a spreadsheet gets read, a PDF gets generated, a manager gets a message.

Most pages ranking for this term are written by CPQ and proposal vendors, and they assume a clean catalog in a modern CRM. Many companies with 50 to 500 people price from an Excel file, an ERP price list, or the head of the person who has done it for twelve years. Sales quotation automation for them starts somewhere else.

Where does the time in a quote actually go?

Time in a quote leaks in four places: intake, pricing, approval and follow-up. Typing the document is rarely the expensive part. Before you buy anything, measure which of the four holds your quotes up, because each one has a different fix.

• Intake. The request arrives as an email, a phone call written on a pad, a PDF drawing, or a WhatsApp message to a rep. Someone has to turn it into line items before any pricing starts. In distribution and manufacturing this step alone can take longer than the rest combined.
• Pricing. The rep finds the right price list, checks customer-specific terms, adds freight, and converts currency. When the price list lives in three versions on three laptops, errors start here.
• Approval. Anything above a discount threshold or below a margin floor waits for a manager. The quote is finished; it is simply not allowed to leave.
• Follow-up. The quote goes out and nobody asks about it again until the buyer has signed elsewhere.

A quick way to find your leak: pull the last 30 quotes and write down four timestamps for each, request received, priced, approved, sent. The longest gap is your project. Most teams that try this are surprised by where it lands.

Approval is the step that stalls, and software alone rarely fixes it.

In most surveys of the problem, the quote is ready long before it is allowed out. Approval is the step that holds it, and the usual cause is a vague approval rule. Automating a vague rule only makes it fail faster.

Two vendor surveys, from companies that both sell quoting software, point the same way. Conga's State of Commercial Operations 2026, covering more than 1,200 decision-makers in manufacturing, financial services, retail and other sectors, found that 45% lost a deal in the past six months because a quote approval was slow, and 93% said deals struggle to move between sales, legal, finance and pricing. In Aleran's manufacturing survey, complex approval processes were the most cited cause of slow quoting, at 53%. Treat both as vendor research with an interest in the answer. Two competitors reaching the same finding from different samples still counts for something.

What a working approval rule looks like

A good approval rule is short enough to fit on a sticky note and specific enough that software can apply it without asking anyone. "Discounts above 12% or margin below 18% go to the sales director; everything else sends automatically" is a rule. "Big deals need a look" is not. A chain built on the second kind sends every quote to the approver, who then spends the day on quotes they would have passed in five seconds.

Where a person should stay in the loop

Keep a human in the loop wherever a wrong quote is expensive to take back: a non-standard discount, a custom spec, a new customer with no credit history. The automation's job there is to deliver the decision with everything attached, the customer's history, the margin, the last three prices they paid, so the approver decides in one message instead of opening four systems. The approver should see a single question and two buttons.

Do you need a CPQ system to automate quotes?

No. CPQ is the right answer when you sell configurable products with many options, your catalog already lives in a CRM, and someone in the company can own the system after launch. When those conditions are missing, a CPQ rollout tends to become a data-cleaning project that runs for months before the first quote goes out faster.

That is probably part of why only 22% of Aleran's manufacturers run CPQ despite the pain they report. A CPQ system needs the catalog, the pricing rules and the approval logic modelled inside it before it produces anything, and the model has to be kept current as prices and products change. For a 120-person distributor whose prices sit in an ERP and an Excel file, the cheaper path is usually to automate around the systems they already have: read the request, pull prices from the ERP, apply the rules in code, generate the document, route the exceptions. Our piece on automating order processing covers the same pattern one step later, once the quote has turned into an order.

Three ways to get it running, with what each costs you

• Buy a CPQ or proposal tool and run it yourselves. You get a system you control and a vendor roadmap. You also get the setup, the catalog modelling and the upkeep, and those land on sales ops or IT for as long as you use it. The rod is excellent; whether it leaves the closet depends on who has the hours.
• Hire an integrator or development shop. They build it, often well, and hand it over. If you want an asset at the end, code or a configured system on your own account, this is the better purchase, and you should choose it. The catch is the fisherman problem: when the project team leaves, so does the knowledge of why each rule is the way it is.
• Have someone build, run and maintain it for you. This is what Uplift does. You describe how quotes work today, in your own words; we build the automation around your existing ERP, spreadsheets and inbox, run it, and keep it working when a price list moves or a system changes. The price is flat, with no seats or usage meters. What you give up is ownership: there is no code or configured system that becomes yours, and once the plan ends the quotes stop running through it. Ask any supplier of this model, us included, what you will have in hand on the day the plan ends, and decide whether that answer is acceptable before you sign.

A skeptic would say the third option is the second one with a subscription attached. The difference is who carries the repair work in year two, when the ERP gets upgraded and the price-list export changes format. With a build-and-hand-over project, that day is a new invoice or a broken quote flow. With a managed model, it is the supplier's problem by design. Whether that is worth giving up ownership depends on how much in-house capacity you really have, which is a question only you can answer.

Should AI write the quote?

AI should read the request, and code should write the numbers. A language model is very good at turning a messy email or a scanned drawing into a list of items and quantities. It is the wrong tool for multiplying a quantity by a price, applying a tiered discount, or adding freight, because a model can produce a confident wrong total and a quote is a commitment.

The arithmetic of chained steps makes the point plainly. If a quote passes through ten steps and each one is right 95% of the time, the whole quote comes out right only 59.9% of the time, because 0.95 to the tenth power is 0.599. Ten steps is not a lot for a quote: read request, match products, find customer, find price list, apply terms, apply discount, add freight, check margin, generate document, send. Every step you move from a model to tested code goes from "usually right" to "right every time it ran the same way yesterday."

That is how we split the work: about 95% tested code and 5% AI, with the AI confined to reading unstructured input such as the customer's email. Prices, totals and approval decisions never pass through a model. The reasoning is laid out on how we build. It also matches the wider record on AI projects: Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027, citing cost, unclear value and weak risk controls. Letting a model price your deals is a fast way to add a risk-control problem.

How fast does a quote need to go out?

Faster than your competitor's, and in most B2B markets that means the same business day for a standard request. No independent study pins down a universal number for quotes specifically, so be suspicious of any page that offers one. What the evidence does show is that speed of response and ease of buying both matter, and that buyers increasingly prefer not to wait for a person.

The most cited research on response speed is old and about leads, not quotes. Harvard Business Review's 2011 study of 1.25 million leads, "The Short Life of Online Sales Leads", found that companies contacting a lead within an hour were nearly seven times as likely to qualify it as those waiting even one hour longer. It measured first contact, not price offers, and buyer behaviour has moved since. It shows which way speed cuts and gives no benchmark for quotes.

The more recent signal is about who buyers want to deal with. In a Gartner survey of 632 B2B buyers, published in a press release on 25 June 2025, 61% said they prefer a buying experience with no sales rep involved. Reps still matter for complex deals. For a standard quote, waiting until the rep is back from a site visit is friction the buyer did not ask for. One vendor dataset points the same way: across 742,137 proposals sent through Proposify in 2025, analyzed in its State of Proposals 2026 report, those with an e-signature closed 60% faster and at a 15% higher rate. It is one company's customer data, so it says more about what is possible than what is typical.

For the wider sales cycle around the quote, routing, CRM updates and handoffs, our guide to sales process automation covers which steps tend to break after launch and why.

Where quote automation stops helping.

Sales quotation automation fixes delay and error in a process that already has rules. It does not fix a process without them, and it does not fix a price nobody will pay. Three situations are worth recognising before you spend money.

Every deal is a negotiation. If the owner sets each price personally after a phone call, there is no rule for software to apply. You can automate intake and follow-up around that person, and you should, but the bottleneck stays human. Write down how they decide, even roughly, before automating the middle.

Volume is low. A company sending eight quotes a month will rarely recover the cost of a full system. A shared price list, an approval rule written down, and a reminder to follow up on day three may get most of the benefit. Automation pays when the same steps repeat often enough that the delays add up to real lost deals.

The quote is fast and still losing. If you pull those 30 quotes and find they all went out the same day and still lost, speed was never the problem. Price, product fit or the relationship is. Aleran's respondents put manual quoting at about 5% of annual revenue, a large figure from a company selling the fix. Only your own lost-deal notes will tell you whether your number is anywhere near it, and no vendor, including us, can answer that from outside your company. For your team shows how sales and operations teams usually find out which of their routines repeat often enough to be worth handing over.