The inquiry-to-proposal tool nobody has sold her

Asked what she wants built, Sara Rojas described the product as a composition of three tools she already pays for: ChatGPT for the language, Canva for the layout, Pinterest for the visual sourcing. One place, driven from her inbox, no coding. It would save the two to three days each large event proposal currently costs her.

The spec, as she gave it

Galina Fendikevich put the shape back to her on the record and she confirmed it: an interface where she connects her inbox, and the system takes an inquiry through to a generated proposal for her to review. No agent configuration, no back-end setup, nothing that requires her to become technical.

Sara added the detail a builder should not skip. "Just put in there like it streams the proposal as it reads everything." Not a template she fills in afterwards. A document that begins assembling from the email that triggered it.

Her verdict on the idea was one word: amazing.

What the system has to do

From her account of the current process, the requirements fall out cleanly:

• Read the inquiry and extract the brief. Client, occasion, colour theme, budget, named items, date.
• Source the items she does not stock. Her example is a sofa. Today that means going through several rental suppliers' catalogues by hand.
• Cost the flowers against the brief. She already does this in ChatGPT — stem counts, seasonal substitutions, per-bouquet breakdown inside a budget.
• Assemble a designed document. This is the part she calls hours of manual work in Canva, and the part no AI tool she uses will touch.
• Hold a slot for vendor prices. Some of the elapsed time is waiting on suppliers, and the proposal cannot finalise without them.

Why the pieces exist and the product does not

Every capability in that list is available somewhere. Language generation is solved for her; she uses it daily. Layout tools exist and she is fluent in one. Visual reference search exists. What does not exist is the thing that runs them in sequence against her actual inbox.

The reason matters for anyone deciding whether to build it. Galina makes the point directly in the interview: a technical route to this exists today. Sara is not going to take it. She has a full plate, she is not going to configure agents and connections, and paying a developer is a significant extra cost on a business two people built from scratch. This is the same owner who did not know ChatGPT had account settings.

So the gap is not capability. It is packaging. The buyer for this product cannot assemble it from parts, which is precisely why she would pay for it assembled.

What it is worth

Two to three days per large event proposal, minus whatever the vendors take to reply. For an owner who is also the finance department and the sole client contact, that is not a productivity gain — it caps how many events she can quote for in a season.

She is also not the only person in the building with an unmet composition request. Her co-owner Luisa Ramirez wants Sintra and Canva merged for the same underlying reason: the AI produces the content, and a second tool has to be opened by hand to make it presentable. Two owners, two workflows, one missing seam between generation and layout.

Where to start if you are building it

The narrow wedge is the layout step, because it is the one both owners named and the one no tool in either stack covers. Everything upstream of it — reading an email, drafting copy, costing flowers — Sara already gets from a $20-a-month chatbot. What she cannot get is the finished artefact.

Build the thing that turns a structured brief into a client-ready document, connect it to an inbox, and the rest of the workflow is already running in her account.

Where this comes from

S2E2: The secret AI strategy this small NYC business uses to save thousands — the full interview with Sara Rojas and Luisa Ramirez. Listen or watch: YouTube, Spotify or Apple Podcasts.