An enquiry lands in a shared inbox. Someone forwards it to sales. Sales asks for missing details. Operations checks availability. A manager reviews pricing. The customer waits while the request moves through people, systems and approvals.
The work to produce the quote may only take a few hours. The delay is usually created by the gaps between those tasks.
For UK SMEs, that gap matters. A quote can take days to reach the customer even when the actual drafting is straightforward. Missing information, unclear ownership and slow sign-off add time before anyone has written a single price. AI, connected systems and automation can reduce that waiting, but only if the process already has sensible rules, current information and human oversight.
Working time is not the same as waiting time
A team may spend 20 minutes reviewing an enquiry, 30 minutes checking details and 45 minutes drafting the quote. The customer still may not see it until two days later.
That is rarely because the quote itself took two days to write. It is usually because the enquiry sat waiting for the next person, the next answer or the next approval.
To spot the delay, track the handovers rather than just the task duration:
- Where does the enquiry pause?
- Who owns the next action?
- What information is still missing?
- Which steps depend on judgement, and which are routine?
Once those pauses are visible, the improvement work becomes more practical. Some delays come from poor capture. Others come from routing. Some need a better source of pricing data. A few are right to stay manual because the decision is too sensitive or too unusual.
Where AI and automation can help at each stage
The enquiry-to-quote process usually breaks into a few distinct stages. Each has different delays and different kinds of support.
| Stage | Typical delay | Possible improvement | Human responsibility |
|---|---|---|---|
| Capture the enquiry | The request is read late or logged manually | AI can classify the enquiry, extract key details and create a task automatically | Check the enquiry is understood correctly and assigned to the right owner |
| Identify missing requirements | Follow-up questions are written ad hoc, so the same gaps are missed | AI can compare the enquiry with a minimum requirements list and suggest clarification questions | Decide which questions matter and avoid over-asking |
| Gather business information | Staff search for pricing, service scope, stock, capacity or terms across different systems | Connected systems can surface approved information in one place | Confirm the information is current and suitable for this customer |
| Prepare a draft quote | Re-keying details creates errors and takes time | Automation can populate a quote template and route it to the right person | Verify scope, pricing logic, exclusions and wording |
| Approve and send | Quotes sit in inboxes or wait for sign-off | Workflow rules can flag overdue approvals and remind the owner | Make the final pricing and commitment decision |
The split is worth keeping clear. AI can read, organise and draft. Automation can move work to the next step and chase it when it stalls. People still need to decide what is being promised, whether the numbers are right and whether the business can really deliver.
A hypothetical commercial maintenance example
Consider a hypothetical commercial maintenance business receiving an email asking for a one-off site visit and a possible service contract.
The message arrives in a general inbox. AI classifies it as a quote request, extracts the customer name, location, requested date and a brief description of the issue, then creates a task for the sales coordinator. It also checks the message against a minimum requirements list and drafts a short clarification email asking for the site type, access arrangements and urgency.
When the customer replies, connected systems pull in approved service descriptions, standard response times and current pricing rules. A draft quote is built from the template, with the right service options and terms already in place. If the request falls outside the standard range, the system flags it for review.
Automation then routes the draft to operations for availability checks and to a manager for pricing approval if the discount or margin falls outside agreed limits. If either step takes too long, reminders go to the owner.
At each point, experienced staff still make the calls that matter:
- Does the scope match what the customer actually needs?
- Are the calculations correct?
- Can the team genuinely deliver within the proposed window?
- Are the terms clear and acceptable?
That balance matters. AI should not invent prices, calculate charges without validated methods, promise delivery dates or send unapproved commitments.
What needs to be in place before AI is useful
AI works best when the quoting process is already reasonably clear. If pricing sits in three spreadsheets, service descriptions are out of date and nobody knows who approves what, automation will only make the confusion faster.
Before adding tools, get the basics in order:
- Clear ownership. Every enquiry needs one named owner, even if several people contribute.
- Minimum enquiry details. Agree the information needed before a quote can be prepared.
- Maintained pricing information. Standard rates, discounts and exceptions need to be current and easy to reach.
- Usable service descriptions. Staff need approved wording for what is included, excluded and optional.
- Approval rules. Decide what can be sent straight away and what must be checked first.
These foundations do more than speed things up. They also improve consistency, because the team is working from the same information rather than personal memory or private shortcuts.
Start with one enquiry type, not the whole process
A broad promise is harder to deliver than a narrow pilot. The better starting point is one common enquiry type that causes regular delays.
Map the current process for that request. Note how long it takes from enquiry to first response, and from receiving complete requirements to sending the quote. Track where work stops, who is waiting, and which details are repeatedly missing.
Then improve one or two bottlenecks only. For example:
- use AI to extract and structure incoming requirements,
- add a simple automation to assign owners and chase overdue tasks,
- or connect approved pricing data to the quote template.
Keep the pilot small enough for the team to learn from it. Review whether quotes are going out faster, whether rework is falling, and whether staff are spending less time chasing information.
That baseline matters. If speed improves but accuracy drops, or if the team has to correct the quote later, the process has not really improved.
Measure speed, accuracy and effort together
The useful measures are operational as well as commercial.
Track:
- Time to first response
- Total time from enquiry to quote
- Time from receiving complete requirements to quote
- Staff time spent preparing and chasing each quote
- Quotes needing correction or rework
- Enquiries left without a clear owner or next action
These figures show where the friction sits. They also stop the team from treating speed as the only target. A quick quote is not useful if it is inaccurate, incomplete or too eager on price, timing or scope.
For many SMEs, the real gain is smoother flow. Fewer enquiries sit unattended. Fewer handovers get lost. Ownership is clearer. Staff spend more time on exceptions that genuinely need experience, instead of chasing basic information.
A faster quote still has to be a good quote
The aim is not to remove people from quoting. It is to remove avoidable waiting.
When the handovers are tighter, the business can respond more quickly without losing control of pricing, scope or commitments. That usually means less chasing, clearer answers for customers and more capacity for the team to handle the enquiries that matter most.
Most businesses know AI matters. Few know where to start. A practical place is the point where a good enquiry turns into a slow quote.
