AI Quote Request Chatbot

Why “How much does it cost?” is so hard to answer online

A visitor types “how much for a new website” or “price to clean a 3,000 square foot office” and expects a number. The honest reply is nearly always “it depends,” which a static page cannot follow up on and a busy owner has no time to explain twenty times a week.

So quote requests arrive thin. The estimator opens a message containing a name, an email address and “please send a quote,” then spends the next few days asking the questions that should have been asked at the start. Some prospects answer. Many have already accepted someone else’s quote by then, often not because it was better but because it arrived first.

A quote request chatbot repairs the front of that process. It treats the visitor’s “how much?” as the opening of a short scoping conversation, asks the questions that genuinely drive your price and hands over a request your team can price without a round of emails. It does not replace your estimator’s judgment, and it does not make up numbers to keep a visitor happy.

Three pricing postures, and who decides which one the bot follows

The first decision in any quote chatbot is what it is permitted to say about price. That decision belongs to the business owner, and we make it explicit in writing before anything is built.

In the third posture, the arithmetic is done by ordinary code working from your price table, not by the AI model. Language models are excellent at conversation and unreliable at exact numbers, so we never let one compose a price on its own. The model gathers the inputs and explains the result; the formula does the math, the same way every time.

Whichever posture you pick, the bot applies it consistently. If you have chosen no figures, a persistent visitor will not talk it into a ballpark, and it will explain why a written quote serves them better than a guess.

Scope questions change with the job

A quote chatbot is only as good as its questions, and the right questions depend entirely on what you sell. We build a scope checklist for each job type you quote, based on what your estimators ask today when they call a prospect back.

Example: commercial cleaning

Type of facility, approximate square footage, number of restrooms, cleaning frequency, preferred hours, specialty work such as floor care, and whether a current provider is under contract.

Example: website redesign

Current site address, roughly how many pages, which features must survive the redesign, any online payments or booking, whether content is being rewritten, and who signs off on the final design.

Example: custom fabrication or millwork

Dimensions, materials, finish, quantity, delivery or installation, drawings or reference photos, and the date the piece is needed by.

The bot recognizes which job type a visitor is describing from their first message and runs the matching checklist, skipping anything they already mentioned in passing. If a request fits none of your categories, it collects a description and contact details and routes it to a person for review rather than forcing it through the wrong questions.

We also agree on a stopping point for each checklist. Asking twelve questions before requesting any contact detail loses people; demanding an email at the very first step loses others. For every job type we separate the answers essential to a first quote from the ones that can wait for the estimator’s follow-up.

Photos, measurements and attachments

For many trades and products, a single photo answers more than ten questions. The chatbot can ask for specific pictures, such as one of the whole wall and one close-up of the damaged area, accept drawings or spreadsheets and take measurements typed in any common format.

• Measurements are converted into one unit system and repeated back for confirmation, so “12 by 14” is never silently read as inches when the customer meant feet.
• Uploads are checked for type and size, kept in secure storage and linked from the request rather than attached to emails that bounce.
• If a visitor cannot take a photo right now, the bot sends a link to add it later and marks the request incomplete until it arrives.
• The AI may describe what a photo shows to help categorize the job, but it does not assess condition, measure from images or price from a picture. Those calls belong to your estimator.

What your estimator opens: the quote-ready brief

Every conversation ends in a brief designed to be priced, not read like a story. It is laid out the same way each time so an estimator can scan it quickly:

1. Job type and one-line summary.
2. Scope answers from the checklist, each in its own field.
3. Attachments with thumbnails and links.
4. Requested timing, recorded exactly as the customer phrased it, with no commitment implied.
5. Anything the bot already told them, including any range or calculator estimate shown.
6. Gaps: required answers still missing, so the estimator knows what to ask first.
7. Contact details and preferred channel.

The brief can arrive as an email, a new deal or opportunity in your CRM, a row in your job-management tool, or all three. Where your quoting software accepts incoming data, the brief can pre-fill a draft quote so the estimator starts from the customer’s own answers instead of a blank screen.

Handling “Just give me a ballpark”

Every quote chatbot meets visitors who want a number before they will answer anything. How it responds decides whether they stay in the conversation.

The two things we avoid are the brush-off (“Please fill in the form and we’ll be in touch”) and the fabricated figure. What works better is a short, specific explanation of what drives price for that kind of job, followed by the two or three questions that would move the visitor from “no idea” to a real quote. If you have approved ranges, this is the moment the bot shares them, conditions included.

Some visitors will still leave, and that is acceptable. A shopper who only wanted a number to compare against is not a lost opportunity if your team is no longer spending time chasing them. The conversation log shows how often it happens, which is useful evidence when you are deciding whether to publish ranges at all.

Other pressure points get the same treatment: discount requests, questions about a competitor’s quote and demands for a guaranteed completion date. The bot acknowledges each one and passes it to a person, because only your business can make those commitments.

After the request is sent

The visitor should leave knowing precisely what happens next. The closing message restates what they asked for, confirms how the quote will reach them and gives the turnaround your business has told us it can realistically meet. We display only the timing you set, and it changes the moment you change it.

Behind the scenes, simple automation keeps requests from going cold on your side:

• A reminder to your team when a request has not been opened within a window you define.
• An optional nudge to the customer if a required photo or answer is still outstanding.
• A status the visitor can check, in setups where quotes are tracked in a system we connect to.

If a job needs someone to see it in person before it can be priced, the conversation can switch to scheduling that visit instead. That path is covered on our AI estimate booking chatbot page.

Which numbers show where quotes are lost

A quote chatbot should be judged on quotes won, not chats opened. We tag each request so you can follow it through your pipeline:

With these stages connected, the leak becomes visible. Plenty of completed requests but few quotes sent points to internal capacity. Many quotes sent but few accepted points to pricing or follow-up, which no chatbot can fix, but which the data at least makes impossible to ignore.

How EVOTECH builds a quote request chatbot

We begin with a free consultation by phone or video to look at how quote requests reach you today, what your estimators ask and which pricing posture fits your business. You then receive a fixed-scope written quote for the build, so the cost of the project is settled before work starts.

From there we write the scope checklists alongside your estimators, configure the pricing rules you approve, connect the brief to your email, CRM or job system and test the whole flow against real past requests before launch. After launch we review the conversation logs with you and adjust questions wherever people stall.

Our team is based in the Houston area, has been in business for over 20 years and holds a 5.0 rating on Google; we build these systems remotely for companies nationwide. The same toolset underpins our wider AI chatbot and AI automation projects. If deciding which requests deserve fast attention is the bigger issue, our AI chatbot for lead qualification page covers scoring and routing.

Related services

Frequently asked questions