The requirements checklist: from a solicitation to one list you work down

Firmenza's fixed rules, not AI, list the sentences that read as requirements, one row each, with the page it came from, the section heading when the document names one, and the document. Each row has a kind, such as Deadline, Submission, Evaluation, Set-aside or registration, Form, Pricing or Work, and reads as required or as preferred: words such as shall, must and required make a row "Reads as required", words such as should and encouraged make it "Reads as preferred", and a due date is always a required deadline. When the deadlines name more than one date, the card says so and asks you to check which applies.

Rules can miss or misread a requirement, so check each row against the document. The card names every page that was not read and every reason a page can go unread, so you know which pages to read yourself. "Correct this row" lets you change a row's words, its kind, and whether it is required or preferred, and "Add a requirement" adds one of your own, marked "Added by you".

Each row has a status you set: it starts at "Needs review", and you move it to "Missing", "In progress", "Complete" or "Not applicable". Nothing is marked complete by software. "Who is on it" names the person handling the row, "Evidence or where it is" holds where your proposal answers it, and a note holds anything else. Words only in each: no tax, bank or account numbers.

"Where this stands" counts the rows that need attention, the complete ones and the not applicable ones, lists the deadlines the document names, and says how many rows that read as required have no evidence yet. "Your details, to compare" shows the fields of your capability statement beside the list, so you can read a requirement and your own words together. Firmenza does not decide whether your business meets a requirement.

The AI assistant reads the same text and lists more. Each of its rows is marked "Listed by the AI assistant" and carries a quote from the document, which Firmenza checks against the page the row names before the row is shown; a row whose quote cannot be matched is left out, and the card says how many were. It reads in the background, the pages where the rules found requirements and Sections L and M, and the list updates when it is done. Where it finds the document's key dates, it proposes each as a task, a card in your inbox that waits for your decision.

It lists what the document says. It does not say whether your business should respond, and it leaves size, set-aside, ownership and certification questions to the document's own words and to you. The card just below this post lists free counselors who can help with those.

Section L holds the instructions for your proposal and Section M the factors it is evaluated on. Where a solicitation has them, the table lists each instruction and each factor in the document's own words with its page, marked "Instruction, Section L" or "Factor, Section M". Firmenza's rules find the headings and make a first split, marked "Found by rules"; the AI assistant splits them finer, each row with its quote.

For each row you write where your proposal answers it, for example Volume II, section 3.1, and mark it "Open" or "Done". Not every solicitation has these sections, and our guide on Sections L and M says which do and where the same information sits otherwise.

"Download CSV" gives you the checklist as a spreadsheet, with each row's kind, page, section, status, who is on it, evidence and note, and "Print" prints the list alone, with the date and Firmenza's disclosure. The Section L and M table has the same two. Both also go into your package, beside your draft and your capability statement; our post on going from a saved notice to a proposal draft shows that path. The pricing page says what each package includes.