Skip to content
Briefing 069 minute readSales and quoting

Draft tender responses from your best past answers

Bid teams spend about 33 hours on each response, and half say they lack the bandwidth. An agent can draft the repeat answers from approved past ones, with the source shown, and leave people the parts that win.

Tom BuckleySpurbuck Advisory
Average share of RFPs won in 2025: 39 in 100Loopio 2026 RFP Trends & Benchmarks

The short version

  • The Department of Finance says the Commonwealth awarded $104.9 billion to suppliers in 2024-25, the highest value ever reported on AusTender, and 45,556 of those contracts went to small and medium businesses 1.
  • Loopio's survey of 1,533 companies that respond to requests for proposal found teams spent an average of 33 hours on each response in 2025 and submitted 166 of them, while the average win rate fell to 39% from 45% 4.
  • Half the respondents in the same survey ranked having enough bandwidth as a top challenge, up from 31% a year earlier, and 49% named working with subject matter experts 4.
  • Much of a response is not new writing: it is the same company description, policies, method statements and project examples, found and adapted again under deadline.
  • An agent can read the request, list every criterion and mandatory form, draft the repeat answers from an approved library with the source of each shown, and hold anything stale or new for a person, so the bid team spends its hours on pricing and win themes.
$104.9bnawarded by the Commonwealth to suppliers in 2024-25, the highest value reported on AusTender 1Department of Finance
33hrsaverage time a team spent on one RFP response in 2025, from receiving it to submitting it 4Loopio 2026 report
39%average share of RFPs won in 2025, down from 45% the year before 4Loopio 2026 report
50%of responders ranking having enough bandwidth as a top challenge, up from 31% 4Loopio 2026 report

Government buys through tenders, and judges what is written

Government buys on a large scale, and most of it through tenders and requests for quote. The Department of Finance reports that in 2024-25 the Commonwealth awarded $104.9 billion to suppliers, the highest value ever reported on AusTender. Of those contracts, 45,556 went to small and medium businesses, 52.4% of all contracts by number 1. Those figures cover the Commonwealth alone; states, councils and government businesses run their own tenders on top.

The Department's own advice to suppliers is blunt about how a response is read. A response that does not meet the conditions for participation will be set aside and excluded from further consideration. Late tenders cannot be accepted. Submissions are evaluated on content, not extravagant presentation, and value for money is the key consideration for Australian Government buyers 2. A tender is won or lost on whether every question is answered, with evidence, by the deadline.

New Zealand's public sector is a large buyer too. New Zealand Government Procurement reports that government agencies spend approximately $51.5 billion a year on goods and services 3. Private buyers issue their own requests for proposal and supplier questionnaires on top.

Where the hours go in a response

Loopio, which sells software for responding to requests for proposal, publishes a yearly benchmark report. Its 2026 report, made with APMP, the bid and proposal professionals' association, and with the data analysed by Qualtrics, surveyed 1,533 companies worldwide, all involved in responding to RFPs; Australian firms were 1% of them, so treat the figures as an international benchmark rather than a local one 4.

Teams spent an average of 33 hours on each response in 2025, measured from receiving the request to submitting it, two hours less than the year before. They submitted an average of 166 responses, up from 153. Loopio's own arithmetic puts that at 5,478 hours a year for the average team 4. Small businesses averaged 27 hours a response and enterprises 39 4.

The hours are not the only strain. Half the respondents ranked having enough bandwidth as one of their biggest challenges, up from 31% a year earlier, and 49% named collaborating with subject matter experts, the challenge that had topped the list every year since 2020 4. Anyone who has run a bid knows the pattern: the operations manager who writes the method statement is also running the operation, and the answer arrives the night before lodgement.

Average share of RFPs won, 2019 to 2025The average win rate reported by respondents to Loopio's survey each year, for the year before each report 4. The 2026 report covers 2025.

More bids is not the answer on its own

The same report found teams submitting more while winning a smaller share. Use of a formal go/no-go process fell to 75% of teams, from 83% the year before, while 81% of the teams Loopio counts as top performers used one 4. Loopio puts part of the rise in submissions down to lower selectivity and heavier use of AI, and 79% of teams said they had used AI in their RFP workflows 4. A tool that only makes it faster to say yes to every request makes this worse.

So the outcome worth paying for is not more tenders. It is the repeat work done in a fraction of the hours, a clear look at each opportunity before anyone commits a week to it, and the time that frees put into the parts a buyer actually scores: the price, the method for this job, and the reasons to choose you.

A tender response by hand

  • Someone reads the whole request to work out what is being asked and what must be attached
  • Past answers are found by searching old submissions, often an out-of-date version
  • Subject experts are asked to write from scratch, late in the timetable
  • Pricing and win themes get what time is left at the end
  • The compliance check happens the night before lodgement

A tender response with an agent

  • Every question, criterion, form and deadline is listed the day the request arrives
  • Repeat answers are drafted from the approved library, with the source of each shown
  • Experts review a short draft of their part instead of starting from nothing
  • The bid team starts on pricing and win themes in the first days
  • Every item on the list is checked against the response before a person lodges it

What we would build

The build is an agent that works inside the places a bid team already uses: the tender portals and inbox where requests arrive, the document library or SharePoint site that holds past submissions, policies, CVs and project sheets, and the CRM where opportunities are tracked. It does not replace the bid manager. It does the reading, the finding and the first drafting, and it shows its working. It does six things.

The tender response agent, line by lineWhat it reads, what it does, what it holds for a person and how you know it worked.
  1. It reads the requestReads

    The request for tender, its schedules and any addenda: every question, evaluation criterion, condition for participation, mandatory form, word or page limit and the lodgement deadline, set out as one checklist.

  2. It summarises the opportunityHolds

    A one-page brief for the bid/no-bid decision: what is being bought, the criteria and their weightings where stated, the mandatory requirements you may not meet, and similar past bids. Whether to bid is the bid lead's decision.

  3. It drafts from the approved libraryDoes

    Each question is matched to approved past answers, and a draft is written for this request, with the source answer and its approval date shown beside it. Nothing is drafted from an answer that has not been approved.

  4. It flags what is stale or missingHolds

    Answers that rely on an old project, a superseded policy, a former staff member or a certificate past its date are held with the reason. Questions with no approved answer go to the named subject expert with what the request asks for.

  5. It checks before lodgementChecks

    The response is checked against the checklist: every question answered, every form and attachment present, limits met. A person reviews the result and lodges the response. The agent never submits anything.

  6. It proves itselfMeasures

    Hours per response, the share of answers drafted from the library and accepted with light edits, items caught at the final check, bid/no-bid decisions made in the first days, and win rate on bids, against a baseline of your last 12 months of tenders.

Everything that goes into the library is approved by a person, and everything that comes out is a draft. Australia's National AI Centre lists maintaining human control as one of its six essential practices for adopting AI, and asks organisations to make sure a person oversees an AI system in a way that matches how much autonomy it has 5. A tender is a promise made to a buyer under your name, so here the agent drafts and a person signs.

What the hours look like for one firm

Take a mid-sized engineering consultancy that responds to 40 tenders a year, a mix of councils, state agencies and private asset owners. A typical response has 62 questions and takes 30 hours, a little under the 33-hour average Loopio measured 4. Four hours go on reading the request and building the compliance list, 12 on finding and adapting past answers, six on chasing subject experts, five on pricing and win themes, and three on formatting and the final check. The firm and its figures are invented; the method is the one any bid manager would use.

With the agent, the bid coordinator checks the agent's compliance list in an hour. Of the 62 questions, 48 are drafted from approved answers, nine go to subject experts as drafts and five are new and written by the bid team. Reviewing the drafted answers takes four hours, the experts spend three, and the final check takes two. Pricing and win themes get six hours instead of five, because that is where the time should go. The response takes 16 hours (1 + 4 + 3 + 6 + 2).

One year of tenders, before and afterIllustrative. The firm, its tenders and every hour are invented, and $110 an hour is a figure chosen for the arithmetic, not a measured cost. Only the 33-hour average comes from a source 4.
Tenders a yearcouncils, agencies, private owners40
Hours by hand40 × 30 hours1,200 h
Hours with the agent drafting40 × 16 hours640 h
Hours back1,200 − 640, or 40 × 14560 h
Hours back, at $110 an hour$61,600

The larger return is not the hours. It is what 560 hours of a bid team's year can do: a proper bid/no-bid look at every request, and more time on the price and method of the bids worth winning.

Watch one tender response, put together two ways

A 30-second film, with no sound to turn on: one 62-question tender answered by hand, then with an agent that drafts from approved answers and holds what is stale for a person.

Illustrative. The company, names and figures are invented for the film.
Read what the film shows
  1. By hand. A 62-question tender takes 30 hours: 4 reading the request and building the compliance list, 12 finding and adapting past answers, 6 chasing subject experts, 5 on pricing and win themes and 3 on formatting and the final check.
  2. With an agent. The agent reads Tender 0471 and lists every question, form and deadline. It drafts 48 answers from the approved library, sends 9 to subject experts and leaves 5 new ones for the bid team. It holds Answer 6.3, whose project example is from 2021 when the buyer asks for work from the last three years, and Sam chooses a 2025 project instead.
  3. What changes. The response takes 16 hours instead of 30, pricing and win themes get 6 hours instead of 5, and across 40 tenders a year the bid team has 560 hours back.

What could go wrong, and how the build guards against it

A confident answer that is no longer true. The most dangerous draft is a fluent one built on an old fact: a project from five years ago, a policy that has been replaced, a manager who has left. The agent drafts only from approved answers, shows each source and its approval date, and holds anything past its review date for a person.

Saying yes to everything. If drafting is easy, the temptation is to bid for more. Loopio found go/no-go use falling as submissions rose and win rates fell 4. The agent's first output is the opportunity brief, and a person decides whether to bid before any drafting starts.

Answers that read like every other bidder's. A buyer reading ten responses notices boilerplate. The library holds the facts; the bid team writes the method and the reasons to choose you for this job, and those parts are never drafted from the library alone.

A missed form or a late lodgement. These lose a tender outright 2. The checklist is built the day the request arrives, the final check runs against it, and the deadline sits in the CRM with the opportunity. A person lodges.

Numbers that prove nothing. Win rates move with the market and the bids you choose. So the build starts from your own last 12 months: hours per response, items found missing at the final check, and win rate by type of buyer. If those do not move, the build has not worked.

Questions for your board

The first three can be answered from last year's bid register this week.

  1. How many tenders did we respond to last year, how many did we win, and what were they worth?
  2. How many hours does a typical response take, and whose hours are they?
  3. Do we decide whether to bid before the work starts, and who makes that call?
  4. Where do our approved answers, policies, CVs and project examples live, and who keeps them current?
  5. Have we ever been excluded for a missing form or a late lodgement?
  6. If our bid team had a quarter of its year back, which bids would we pursue harder?

Run a professional services firm?

Our page on professional services shows the systems we connect, from practice management and documents to finance and the CRM, and the work an agent can take on, from intake to partner reporting.

See AI for professional services

Sources

  1. Department of Finance, Statistics on Australian Government procurement contracts (2024-25, AusTender data as at 13 August 2025)
  2. Department of Finance, Selling to Government: How do I give my tender response the best chance of success?
  3. New Zealand Government Procurement, Government procurement insights
  4. Loopio, 2026 RFP Response Trends & Benchmarks Report (with APMP; survey of 1,533 companies)
  5. National Artificial Intelligence Centre, Guidance for AI Adoption: Foundations (October 2025)

Questions people ask us

How long does it take to respond to a tender or RFP?

Loopio's 2026 RFP Response Trends and Benchmarks Report, a survey of 1,533 companies, found teams spent an average of 33 hours on each response in 2025, from receiving the request to submitting it. Small businesses averaged 27 hours and enterprises 39. Teams submitted an average of 166 responses in the year.

Can AI write tender responses?

An AI agent can draft the repeat parts of a tender response, such as company information, policies, method statements and project examples, from a library of approved past answers, with the source of each shown. A person should decide whether to bid, write the pricing and win themes, review every draft and lodge the response. The agent should never submit anything itself.

How much does the Australian Government spend through tenders?

The Department of Finance reports that the Commonwealth awarded $104.9 billion to suppliers in 2024-25, the highest value ever reported on AusTender, and that 45,556 contracts, or 52.4% by number, went to small and medium businesses. State, local government and private tenders are in addition.

Start with a free discovery session.

Thirty minutes with Tom on how your tenders are put together today, where the hours go, and what an agent would draft and hold for your bid team. You leave with a clear view of what is worth automating, what it would take, and what it would return. No obligation.

We use the details you give us only to arrange and hold your session. See our privacy policy.