Mercha cuts proposal drafting from hours to seconds

Mercha has deployed Decidr OS to automate the creation of custom sales proposals, cutting a process that once took two to three hours down to seconds.

The Sydney-based branded merchandise company previously had to select products, brief an offshore design team on mock-ups, set tiered pricing and format the final document for each proposal.

That workflow is significant because the business handles about 200 instant quote requests a month. With annual revenue of AUD $5.5 million to AUD $6 million and a team of 12, the manual approach limited how many tailored responses the sales team could send to prospective customers.

The new system sits within Mercha's sales workflow and retains a human review stage before any material reaches a prospect. It enriches a lead with company revenue, size and industry data, then assembles a proposal with product recommendations, quantities, pricing and branded mock-ups.

Mercha sells more than 2,000 branded products through a real-time transaction platform. Proposal content is drawn from patterns across similar customers, while mock-up quality and product logic are being refined through repeated iterations.

Ben Read, chief executive officer of Mercha, said the company had focused its use of artificial intelligence on a specific sales task with a clear return.

"We're a high-growth merch-tech business scaling revenue without adding headcount. We mapped every function - ops, marketing, sales, fulfilment - and focused AI where it drives the clearest ROI. Our 200+ monthly instant quotes used to go out as basic pricing emails. Turning them into rich, tailored proposals materially shifts conversion. Our work with Decidr will take that process from hours to seconds. The impact is that a 12-person team can now drastically reduce manual work to focus on what truly matters, delighting our valued customers," Read said.

Lean business

For Decidr, the deployment illustrates how smaller companies are using agent-based software first for narrow, repetitive tasks rather than broad transformations. Mercha has joined as a foundation customer of Decidr OS.

David Brudenell, co-chief executive officer of Decidr, described the Mercha workflow as a clear fit for automation.

"Mercha is exactly the kind of business our platform was built for. Twelve people, thousands of leads, a defined high-repetition workflow where the return on automation is unambiguous. Ben's team identified the most expensive process in their sales motion and rebuilt it inside Decidr's agentic platform. As AI models continue to commoditise, the durable value accrues to the orchestration layer: the infrastructure that turns AI outputs into auditable, real-world business execution. Mercha is a clear example of what that looks like at the lean end of the Australian market," Brudenell said.

The software group was founded in 2018 by Paul Chan. Its system is used in business workflows including hiring, eCommerce conversion, staff onboarding and learning content. Customers include CareerOne, Go1, ELMO, Edible Beauty, Growth Faculty and AIM.

Next steps

Mercha has identified further uses for the system beyond the current quoting process. Its existing instant quote email is set to be replaced by a fully generated proposal, and it is also planning a natural-language quoting experience on its website that would let a customer describe an event in plain English and receive a complete proposal.

Back-office work is also under review. Mercha has identified a 167-step workflow covering supplier coordination, artwork approvals, pre-production proof sign-offs and invoice management, all of which is currently handled by email.

Read said the company's approach was to start with a tightly defined problem rather than wait for a broader overhaul.

"If you believe AI is going to transform every part of every business - and it will - you need a partner building for that future rather than bolting tools onto the old one. Taking no step because you think you need to walk 100 steps first is a recipe for failure. Pick a tightly defined problem, fix it, then expand," Read said.