Mapping a quoting workflow before automating it

What was stuck

We’ve worked with this marketplace since 2020. By 2025, the next stage of growth depended on quoting. A buyer’s question could arrive by web form, live chat, phone or email. From there it moved through a spreadsheet standing in for a CRM, Shopify draft orders, a Drive folder of vendor prices, emails to vendors, an Excel margin sheet and a PDF quote tool.

What we found

In two working sessions in May 2025, we mapped the path from a stranger’s first question to a sent quote, with the general manager and the sales rep who do the work. The same details were typed in more than once. No one could see where leads came from. Vendor and freight lookups took 24 to 48 hours.

What we built, step by step

1. February 2025. Scenario planning with the general manager, built on a pro forma, with a guardrail for how much of each month’s earnings to reinvest.
2. May 2025. The end-to-end map of quoting and ordering, checked with the people who use it every day.
3. Then the future state. A design for six agents, each handling one part of the path, with set points where a person reviews and approves before anything reaches a customer.
4. September 2025. A 15-day review that narrowed the strategy to one market: data-centre buyers.
5. 2026. The owner agreed to the data-centre focus. A new operations lead now runs daily AI routines, including a daily revenue tracker and a weekly performance review, and we check them for data integrity. The email platform is connected.

What changed

The quoting path is written down, checked by the people who use it, and designed with clear approval points. The business has one focus market, and its daily numbers come from routines rather than someone’s spreadsheet.

What isn’t finished: reconciling the order system with the storefront, and building the agents themselves. On the Agentic Ascent, this business is at level 1, Connected, and working on level 2, Reconciled.