The Death of the RFP as We Knew It
Buyers use AI to write the questions, vendors use AI to draft the answers, and the ritual that governed enterprise sales for a generation is quietly ending.
Sign up for my Substack daily AI newsletter here.
See my AI Training course portfolio for corporate Business Leaders here.
Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/
The Request for Proposal (RFP), that long, unstructured, chased-by-phone-calls document that decides which vendors win multi-year contracts, is being dismantled at every stage. Buyers now use artificial intelligence to draft the questions, sellers use AI to compose the answers, and buyers use AI again to summarise, score and rank the responses before a human reads them. Recent research finds that ninety-four per cent of B2B buyers now use large language models during their evaluations, and the same share rank their shortlist in order of preference before speaking to any seller. For chief revenue officers and their sales teams, this is not a distant technology story. The first reader of every proposal you now send is a machine, and the way you write for it will decide whether a person ever reads a word.
CONTEXT AND BACKGROUND
The evidence for the shift is stark. In a study of nearly four thousand B2B buyers, the research firm 6sense found that ninety-four per cent use large language models during a purchase, and ninety-four per cent of buying groups now compile a ranked shortlist before contacting any potential vendor. The preferred vendor on that shortlist goes on to win the deal roughly eighty per cent of the time. Gartner reports that seventy-five per cent of B2B buyers now prefer a sales experience with no human representative at all, at least in the early stages.
On the seller side, the shift runs the other way. Response teams use AI to parse incoming RFPs, retrieve content from Salesforce, SharePoint and Notion, draft answers and check compliance, with turnaround now measured in hours rather than days. Buyer procurement platforms, in turn, auto-generate questionnaires, normalise supplier responses and rank offers by weighted criteria before a person opens the file. Almost every step of the RFP that used to be done by a person has been replaced by a model.
INSIGHT AND ANALYSIS
The uncomfortable truth for sellers is that a proposal is now written for two audiences at once, and the machine sees it first. Just as search-engine optimisation once forced copywriters to write for Google’s crawlers before human readers, sales teams are now writing for procurement algorithms before human buyers, and most have not yet noticed. Long, persuasive prose is exactly the kind of writing an LLM strips out when it summarises. Marketing pages built for a human eye may be misrepresented or omitted when a buyer-side agent extracts vendor comparisons. Even carefully crafted case studies lose their emotional weight when compressed into a scoring rubric. HubSpot’s analysis of the shift argues that answer-first formatting, tight paragraphs, verifiable claims and consistent facts across every touchpoint are now table stakes for content read by a machine before a person.
McKinsey’s most recent research adds that growth champions are not bolting AI onto old workflows but rewiring the sales motion around it, from account planning to proposal drafting. The firms achieving five to eight per cent revenue uplift are those that treat the AI-enabled buyer as a genuine change, not a channel adjustment.
IMPLICATIONS
For South African revenue teams, the implications sit close to the ground. Long, discursive proposals need to be shorter and more structured, with numbers, dates and specifications a machine can extract and compare. Product pages and case studies must be written to be quoted accurately by a summariser, not merely admired by a browser. Ambiguity is now an ambiguity tax: what an LLM cannot extract, it scores as zero or omits from the shortlist entirely. Pricing that lives only in a salesperson’s head, deliberately withheld to force a discovery meeting, cannot compete with a rival whose numbers a buyer’s AI can pull cleanly.
The Gartner data adds a useful caveat, though. Buyers who go fully rep-free are more likely to regret the purchase, and hybrid deals combining digital and human interaction produce measurably better outcomes. The sales team’s real task is now twofold: winning the shortlist against a machine, and winning the deal against another human. This is cognitive governance applied to the revenue function. Both skills are required. Few teams are trained for either.
CLOSING TAKEAWAY
The death of the RFP is not the death of enterprise sales. It is the death of the sales veneer. The machine strips away the charm offensive to see whether the product actually fits, then hands what remains to a human committee that must still choose, argue, sponsor and sign. Sales leaders who read this only as a technology story will optimise their proposals and lose. Those who read it as a shift in the audience will rewire how their teams write, market and sell, and will still be in the room when a person finally decides. The proposal you send this week will be read by a machine. Write for it.
Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net
Comments