Shuichiro Ogawa
日本語

Notes · updated 2026-10-07

Consulting Proposals and Competitive Bids: Their Forms, Why Buyers Award and Sellers Win, and What Generative AI Changes (from Industry Sources)

Drawing on industry sources (procurement rules and public audits in the UK, the US, and Japan, industry surveys of RFP response, and statements by consulting executives on earnings calls), this note sorts out the forms of consulting proposals and the reasons why buyers award and sellers win competitive bids, and examin…

Contents (7)
  1. Proposals are written faster; are they chosen more often?
  2. The buyer’s evaluation framework sets the form of a proposal
  3. Buyers and sellers see the reasons for awards differently
  4. What moved after 2024
  5. What generative AI changes
  6. Where this account does not apply
  7. Footnotes

Proposals are written faster; are they chosen more often?

This note is about industry practice. The sources are procurement rules and public audit reports from the UK, the US, and Japan, industry surveys of how firms respond to requests for proposals (RFPs), and statements that consulting executives made on earnings calls and in interviews. Academic research was not collected1.

Generative AI entered proposal writing quickly from 2024. In the annual survey that Loopio, a vendor of RFP response software, runs with APMP (the professional association for proposal management), 34% of teams had tried generative AI in 2023 (Loopio 2024). The share using it rose to 68% in 2024 (Loopio 2025) and to 79% in 2025 (Loopio 2026).

In the same survey, the average win rate rose from 43% in 2023 to 45% in 2024 and then fell to 39% in 2025 (Loopio 2025; Loopio 2026). Loopio attributes the fall to economic conditions and does not examine any causal link with AI. However, Loopio itself notes that few teams name improved win rates as a benefit of AI (Loopio 2026). Writing a proposal faster and being chosen appear to be separate matters.

What, then, has decided which consulting proposals are chosen? Which parts of the forms and reasons moved after 2024, and which did not?

The buyer’s evaluation framework sets the form of a proposal

The chapter structure of a proposal differs from firm to firm. However, what a competitive proposal contains is set by the buyer’s evaluation framework before the seller writes anything. Public procurement documents state that framework explicitly.

In the UK, the Procurement Act 2023 set the award standard as the “most advantageous tender” (MAT) (UK Parliament 2023). The MAT is the tender that satisfies the authority’s requirements and best satisfies the award criteria. The Cabinet Office guidance states that awarding on lowest price only “is unlikely to be suitable for most contracts” (Cabinet Office 2026). Under the government’s joint framework for consultancy (MCF3, 2021 to 2025), further competition among framework suppliers is the default route to market (Crown Commercial Service 2021). Quality is weighted at 50 to 90% and price at 10 to 50%. The examples of quality sub-criteria are approach, solution, resourcing, and mobilization.

The US Federal Acquisition Regulation distinguishes two processes. In the tradeoff process, a contract can be awarded to a proposal other than the lowest priced one, provided that the benefits of the higher priced proposal merit the additional cost and the rationale is documented (GSA et al. n.d.a). In the lowest price technically acceptable (LPTA) process, the lowest priced proposal that meets the requirements is chosen, and tradeoffs are not permitted (GSA et al. n.d.b). For contracts of $10 million or more for IT and support services awarded by the military departments in the first half of fiscal year 2017, LPTA was rarely used (GAO 2017).

For design and consulting services in Japan, the proposal method selects “the technically most suitable” provider on the basis of a technical proposal (National Conference of Government Building Directors 2018). The buyer sets one to three evaluation themes that reflect the priorities of the project, and applicants describe their policy and approach for the work and their technical proposal on those themes. Because the technically most suitable provider cannot be judged by price, the guidance states that it is not appropriate to ask for a fee and use it in the evaluation. The number of pages is limited, and no concrete design is requested. In the Digital Agency’s procurement in FY2023, comprehensive evaluation accounted for 55% of the 150 open competitive contracts by number and 97% by value (Digital Agency 2025a). Separately from open competition, there were 95 contracts awarded through proposal-based competition (25% of all contracts by number).

When the three countries’ frameworks are placed side by side, what a competitive proposal has to describe is common to all of them: how the work will be done and by what method, who will do it and in what team, experience with similar work, and price. This set of contents did not change in 2024.

The pricing models were also in place before 2024. The UK government’s 2022 playbook for procuring consultants (the Consultancy Playbook) lists three models (Cabinet Office 2022). Under time and materials (T&M), the buyer pays for actual time worked according to a rate card of day rates by consultant grade. Under a fixed price, the buyer pays a set amount for delivering the specification within the timeframe, and the supplier arrives at the amount by estimating the work and adding contingency. Under payment by results, part of the fee is charged on a T&M or fixed-price basis, and the remainder depends on the results delivered. The playbook recommends T&M at the start of a project, when scope is unclear, and a fixed price once the requirement is more mature. It limits payment by results to cases in which the outcome can be quantified, a baseline can be agreed, and the consultants possess all the levers needed to realize the outcome. The same playbook also asks buyers to set a quality threshold so that suppliers cannot bypass quality requirements by offering the lowest average day rate.

Buyers and sellers see the reasons for awards differently

The evaluation framework does not capture everything that leads a buyer to choose. The Hinge Research Institute surveyed 822 buyers and 533 providers of professional services and published the results in 2013 (Hinge Research Institute 2013a; 2013b). When searching for a new firm, buyers turned first to friends and colleagues 71% of the time. In the initial selection, a good reputation was the most important criterion at 21%, above an existing personal relationship (18%) and team expertise (12%). Reputation remained most important in the final selection, and the report states that “cost is a minor consideration in the final selection.” Sellers greatly overestimated the importance of expertise and experience. Because the buyers in this study were clients of the sellers, the results lean toward existing relationships and should be read with that in mind.

The reasons sellers give for losing differ from this picture. In the Loopio survey, price has been the top reason for losing a bid every year since 2021 (Loopio 2026). In 2023, 67% said they lost on price, up from 55% two years earlier (Loopio 2024). In 2025, only 13% named proposal quality as a reason (Loopio 2026). Because Hinge and Loopio differ in year and population, their figures cannot be compared. Nevertheless, the pattern in which buyers choose on reputation and relationships while losing sellers blame price points in the same direction as the sellers’ misperception that Hinge describes.

Public-sector buyers moved after 2024 toward requiring a justification for using outside consultants at all. According to the UK National Audit Office (NAO), central government spending on consultants was about £1.36 billion in 2022-23, and the Chancellor announced the intention to halve that spending in 2025-26 (National Audit Office 2025). Among its lessons, the NAO lists ensuring that contracts are focused on outcomes and outputs rather than inputs (people and time). In the same report, 86% of officials surveyed in departments and arm’s-length bodies rated consultants as valuable. In April 2025, the US General Services Administration (GSA) announced stricter controls that require agencies to defend all external consulting spend (GSA 2025a). These cuts come from the governments’ spending policies, and none of the documents names generative AI as their cause.

What moved after 2024

The work of writing proposals

The spread of generative AI is shown by the Loopio figures at the start of this note. In 2025, 84% of users used AI at least once a week, and the most common uses were generating specific answers, editing, and writing first drafts, in that order (Loopio 2026). For the time spent per proposal, the 2025 report says that writing time fell from 30 hours to 25 hours (Loopio 2025), and the 2026 report says that teams spent 33 hours from start to finish, two hours less than the year before (Loopio 2026). The two figures measure different things and cannot be joined, but both point toward shorter times. In 2025, management consulting overtook insurance as the industry with the most annual submissions (Loopio 2026). However, management consulting accounts for only 5% of Loopio’s respondents.

The buyers’ rules

Buyers did not prohibit suppliers from using AI; instead, they moved toward asking for disclosure. The UK Cabinet Office’s Procurement Policy Note PPN 02/24 (March 25, 2024) acknowledges that suppliers using AI to develop their bids can “bid for a greater number of public contracts” and states that suppliers’ use of AI is not prohibited (Cabinet Office 2024). It also gives example questions asking suppliers to disclose their use of AI in creating their tenders. The revised version, PPN 017 (February 2025), keeps the same position (Cabinet Office 2025). The US Office of Management and Budget (OMB) memorandum M-25-22 (April 3, 2025) asks agencies to determine whether circumstances merit including a provision in a solicitation that requires disclosure of AI use in contract performance (OMB 2025).

Buyers have also begun to use AI themselves. In August 2025, the GSA asked industry for input on a procurement ecosystem that incorporates AI (GSA 2025b). In a July 2025 document, Japan’s Digital Agency listed among its next steps a study of drafting procurement specifications with generative AI and a review of the conditions for applying a 1:3 ratio of price points to technical points in comprehensive evaluation (Digital Agency 2025b). In May 2025, the agency issued a guideline on how the government procures and uses generative AI (Digital Agency 2025c). All of these are at the stage of study and policy, and no results yet show how the evaluation of proposals has changed.

The mix of pricing models

At large firms, the mix of the three pricing models is shifting. On the December 2025 earnings call, Accenture’s CFO said that about 60% of its work in FY25 was fixed-price, up about 10 points over the last three years (Accenture 2025b). The CFO attributed this to the increasing role of the firm’s proprietary platforms over a long period and to clients wanting greater certainty in cost and delivery. On the same call, the CEO said that the decoupling of revenue from headcount had been going on for a long time and went back to the introduction of RPA. On the call three months earlier, the CFO said that pricing on advanced AI work was accretive to the firm’s average, and the CEO explained that money clients save through AI frees up their budget for their next priorities (Accenture 2025a).

On its call for the fourth quarter of 2025, Cognizant said that fixed-bid and transaction-based work now represented more than 50% of its revenue, and that clients were re-baselining their discretionary spending in expectation of productivity gains (Cognizant 2026a). On its July 2026 call, it said that some clients had started to ask for an AI-infused rate card that embeds pretraining and inference costs (Cognizant 2026b). The CFO explained that T&M work, which is renewed on short cycles of about six to nine months, embeds the benefit of AI each time it is renewed.

In strategy consulting, executives talk about fees linked to outcomes. In January 2026, McKinsey’s global managing partner said that about a third of the firm’s revenue underwrites outcomes and expressed the hope that this would cross a majority of revenue by the end of the current term as global managing partner (Harvard Business Review 2026). In the same interview, the global managing partner described the firm’s workforce as “40,000 humans and 20,000 agents.” BCG’s CEO said that three quarters of the firm’s largest AI cases now have variable-fee arrangements, while the share across all of BCG’s work is significantly less than a third (BCG n.d.). The CEO added that this is not new, since the CEO’s very first case as a partner had a variable fee. Neither firm has disclosed how it defines outcome-linked work.

In advertising, WPP’s CEO said in August 2026 that the time and materials model is probably not sustainable in the long term, because AI will let the agency do its work faster with fewer people (Bradley 2026). However, only one client has so far adopted outcome-based pay, and the CEO expected the shift to take a few years. In Japan, Nomura Research Institute said at its April 2026 briefing that all revenue in its consulting segment would come to be AI-related and that it wanted to raise unit prices for AI-related work (Nomura Research Institute 2026). It added that AI-related projects were growing larger, with some reaching several billion yen, while noting that defining AI-related revenue is “very difficult.”

On the buyer side, an IDC survey of 72 CIOs and CTOs in October 2025 found that consulting and systems integration was the category of IT vendor that respondents were most likely to change in the next year, at nearly 20% (IDC 2025). IDC concludes that the link to business outcomes is the most critical factor for preserving the value of consulting and systems integration services. IDC itself notes that the sample is small.

What generative AI changes

The sources above support three hypotheses about what changes. None of them has been measured directly; at this stage, several sources merely point in the same direction.

First, the gap between proposal documents narrows, and the weight of selection moves outside the document. If writing costs less, a seller can bid on more opportunities with the same staff. The UK Cabinet Office describes this as a benefit of AI (Cabinet Office 2024). If every bidder submits a polished proposal with the same tools, the quality of the document is less likely to separate them. The fact that win rates have not risen (Loopio 2026) is consistent with this view. In that situation, buyers rely on what already weighed heavily, namely reputation and existing relationships (Hinge Research Institute 2013b), and on the presentations and interviews that the UK playbook recommends (Cabinet Office 2022) and the hearings in Japan’s proposal method (National Conference of Government Building Directors 2018). This hypothesis would be wrong if buyer-side data showed that, after the spread of AI, scores on the proposal document remain decisive in buyers’ evaluations.

Second, the basis for pricing by time weakens, and who takes the share of productivity gains becomes the subject of negotiation. A T&M rate card sets the price by day rates per grade and the number of days (Cabinet Office 2022). If AI reduces the days needed for the same deliverable, the reason to pay by the day weakens. Cognizant’s clients have lowered their spending baselines in expectation of productivity gains and are asking for rate cards that include AI costs (Cognizant 2026a; 2026b). On the seller side, the share of fixed-price work (Accenture 2025b) and of outcome-linked work (Harvard Business Review 2026) is rising. Under these models, the price is not proportional to days, so the days saved through AI can remain in the seller’s margin. Accenture’s statement that AI work is priced above its average (Accenture 2025a) and Nomura Research Institute’s wish to raise unit prices (Nomura Research Institute 2026) show the seller’s stance in this negotiation.

Third, payment by results spreads, but only to engagements with measurable outcomes. As early as 2022, the UK playbook limited payment by results to cases in which the outcome can be quantified, a baseline can be agreed, and the supplier holds all the levers (Cabinet Office 2022). At BCG, variable fees are concentrated in the largest AI cases and remain significantly below a third of all work (BCG n.d.). At WPP, only one client has agreed to such fees (Bradley 2026). Engagements that can promise outcomes that are easy to count, such as cost reduction or processing speed, are likely to move to payment by results first. In engagements such as strategic advice, where the outcome depends on factors beyond the seller’s reach, T&M or fixed prices will probably remain.

Some things do not change. The evaluation frameworks (the UK MAT and its quality weighting, the US tradeoff, and Japan’s proposal method) are the pre-2024 ones still in use. The three pricing models were already in the 2022 playbook, and BCG’s CEO says that variable fees are not new. According to Accenture’s CEO, the decoupling of revenue from headcount has continued since the introduction of RPA (Accenture 2025b). What generative AI changes is less the forms themselves than the weights and proportions within them.

Where this account does not apply

Most of the RFP figures come from the Loopio survey, whose respondents are proposal staff on the seller side, of whom 5 to 7% are in management consulting and most are in North America. Win rates and reasons for losing are self-reported. The only source on buyers’ selection reasons was the Hinge study published in 2013, and no source measuring buyers’ selection reasons after 2024 with a stated method was found.

No public statistics measuring the effect of generative AI on the number of bids or the reasons for awards were found either. The Digital Agency’s share of single-bidder contracts fell from 47% in FY2022 to 38% in FY2023 (Digital Agency 2025a), but the agency explains this as a result of its own measures to increase competition and does not connect it with AI.

The figures on firms’ pricing models are executives’ statements about their own firms, and their definitions differ. McKinsey’s “about a third,” BCG’s “significantly less than a third,” and the fixed-price shares of Accenture and Cognizant count different things and cannot be compared directly.

For the proposal and pricing practices of private consulting firms in Japan, no source other than the earnings briefing statements (Nomura Research Institute 2026) could be obtained. Pitches in advertising and design were not covered, because the full texts of the industry associations’ surveys are restricted to members.

Unverified items

No unverified statement remains in the main claims. Sources that could not be reached (GSA’s CALC+ labor rate data, the evaluation weights of the UK MCF4, Source Global Research’s buyer survey, the ANA and 4A’s survey of pitch costs, and Japan’s national single-bidder rate) are recorded in the ledger with the routes attempted.

References

Footnotes

  1. The sources are sorted into three tiers: public bodies and regulators (T1), the factual parts of surveys by research firms and vendors (T2), and statements by company insiders (T3, developer voice). Loopio, which sells RFP response software, and executives describing their own pricing models were treated as sources that may contain position talk, and only their figures and the fact of their statements were used. Secondary news summaries were not used. The ledger is at source/review/consulting-proposal-procurement-genai/industry.md. ↩

Author: Shuichiro Ogawa (Design Researcher / Consultant) About me →