Imagine you lead a national equipment distributor, with regional sales teams and distribution centres across the country. Your sales teams regularly receive requests for equipment quotes from construction companies.
Let’s work through one of those enquiries as an example. A construction company reaches out and says it needs generators for several building sites.
Before your sales team can prepare a quote for this customer, it needs to confirm how much the order will cost and when it can be delivered. To do that, the team gets the customer’s contract prices from central pricing and asks operations to check stock across the distribution centres. If there aren’t enough generators available, sales then checks with the manufacturer to find out when the rest can be supplied.
While your teams collect those answers, the construction company is waiting for your quote to decide whether to order from you or a competitor.
Now let’s look at how an AI-enabled workflow could gather those answers and prepare a quote for your sales team to review. Our aim is to respond sooner and give your team a better chance of winning the order.
Bringing the pricing and stock checks together
Our proposed workflow starts with two steps:
- Understand what the customer wants: AI reads your customer’s enquiry directly so it can identify which generators the customer needs, how many they want and where they want them delivered.
- Check what the distributor can supply: The AI workflow then checks the customer’s contract prices, how many generators are available and whether the distributor can deliver to their requested sites. Pricing and stock checks also run in parallel.
To check a price, the workflow uses a software connector to retrieve the customer’s contract price from the pricing system.
Model Context Protocol, or MCP, sets a standard for how AI applications use these connectors. If a suitable MCP connector already exists for your pricing system, your team can reuse it to add the price check to the workflow with less custom code.
The stock check also needs to exclude generators already reserved for other customers. That tells sales how much of this customer’s request can be fulfilled from available stock.
Finding delivery dates for the remaining generators
If that stock check shows a shortfall, the next step is to find out when the manufacturer can supply the remaining generators.
Your sales team might normally check the manufacturer’s portal for that delivery date. If the portal has no suitable software connection, a computer-use agent could retrieve the date by reading and interacting with the portal’s screens.
I’d test that step on the actual portal before relying on it. If the date is missing or unclear, the agent needs to flag the enquiry for sales to follow up with the manufacturer.
Getting approval before the quote goes out
Once pricing and delivery details are confirmed, the workflow can prepare a quote for sales to review. If sales wants to offer a discount outside the customer’s contract, it needs the sales manager’s approval.
The workflow would handle that approval in three steps:
- Send the quote for review: Give the manager the proposed quote and the pricing, stock and delivery information for it.
- Wait for the decision: Save the quote and the workflow’s progress so the process can resume when the manager responds.
- Recheck before sending: After approval, check that prices and availability still hold. Flag any changes for sales to resolve before the quote goes to the customer.
Would this justify the investment?
To test whether investing in this approach would pay off, I’d compare a pilot with the current process. I’d measure the time from the customer’s enquiry to a quote with confirmed prices and delivery dates, alongside the volume of enquiries the sales teams handle. Then I’d look at whether the workflow helps them handle more enquiries and win more orders.
Those improvements would need to justify the cost of connecting the systems, running the workflow and handling exceptions.
For a government agency processing grant applications, a similar workflow could gather records from several systems into case files for assessment officers. I’d measure the time officers spend collecting records, how many applications they assess, and the number still awaiting a decision.
Across many teams, those improvements could mean more business won or shorter waits for applicants. Those outcomes would make a strong case for the investment.