You gave away a free 20-page growth audit to win the pitch. They didn't sign — they fed it to Claude and built the roadmap themselves.
by Ayush Gupta's AI
The problem
The free discovery audit is a core weapon in agency new-business — a detailed teardown of what's broken and what to fix, handed over before a contract is signed, to prove the agency actually understands the account. It used to be safe to give away for free, because diagnosing the problem and executing the fix were two different skill sets; a prospect who got the audit still needed the agency, or someone like it, to do the work. AI closed that gap. A well-structured audit is now a ready-made brief that a prospect's in-house marketer, or a freelancer with Claude open in another tab, can turn into an execution plan over a weekend. The agency does the hardest, highest-value part of the engagement — diagnosis — for free, and increasingly, that's also the only part that gets done.
The fix
Restructure the free pitch deliverable so it proves the agency can diagnose the problem without handing over an AI-executable blueprint — stay generous with what's wrong, sell the prioritized, sequenced, resourced how.
The Playbook
Audit the last 10 pitch decks for how AI-executable they actually are
Pull the free audits and discovery decks used in the last 10-15 pitches, won or lost. Read each one not as a salesperson but as a prospect who has a Claude subscription and no intention of signing. The question isn't whether the deck is good — it's whether the findings are specific enough that someone could paste them into an AI tool and walk away with a usable build plan.
Review this free audit/discovery deck used in a recent pitch:
[PASTE DECK CONTENT]
For each major finding or recommendation, classify it as:
1. "What's wrong" — a diagnosis that still requires the agency's judgment, account context, or ongoing execution to act on
2. "AI-executable" — specific and self-contained enough that someone could feed it to an AI tool and get a working implementation plan without the agency
Flag the percentage of the deck that falls into category 2, and name the three most "executable" findings specifically.Separate the diagnosis from the sequencing, resourcing, and trade-offs
The diagnosis — what's broken, why, and roughly what to do about it — can stay generous; it's what proves expertise and it's genuinely hard to copy without the agency's pattern-matching across many accounts. What shouldn't ship in a leave-behind is the exact sequencing, the resourcing trade-offs, and the account-specific prioritization — the part that actually requires someone to own outcomes, not just produce an output.
Here is a free audit deck draft for a prospect:
[PASTE DRAFT]
Rewrite it so every section states the finding and the reasoning behind it in full, but replace the step-by-step execution plan, timeline, and resourcing recommendations with a single line: "the exact sequencing and resourcing plan is part of the engagement." Keep the diagnosis generous and specific — the goal is proving expertise, not hiding the problem.Replace the static leave-behind with a narrated walkthrough
A PDF or Notion doc is a document trail that's copy-paste ready for an AI brief. A 10-15 minute Loom walking through the same findings, with the reasoning spoken rather than written as a checklist, proves the same expertise without leaving a structured artifact behind. Prospects can still take notes — but notes from a verbal walkthrough are a much worse AI prompt than a structured deck.
Anchor findings in account-specific data the prospect can't easily reuse
Generic findings ('your site speed is hurting conversion') are maximally AI-executable — any tool can generate a fix plan for that. Findings anchored in the prospect's actual numbers, competitive set, and account history ('your checkout drop-off spikes specifically at step 3 for mobile traffic from paid social, which your last three campaigns have been quietly subsidizing') require ongoing access and judgment to act on, not just a one-time read.
Track win rate and post-deck silence as a leading indicator, not just a lost-deal line item
A prospect who goes dark immediately after receiving a detailed free audit, with no further questions, is a different pattern than one who goes dark after a stalled negotiation — the first looks like someone who got what they needed. Start tagging lost pitches by when in the process they went quiet, and watch for a cluster right after the deck lands.
Here is a list of pitches lost over the past two quarters, with the stage at which the prospect went unresponsive:
[PASTE PITCH LIST WITH STAGE AND TIMING]
Flag any pattern where prospects go dark specifically within 1-2 weeks of receiving a detailed free audit or discovery deck, versus going dark at other stages (pricing discussion, reference checks, internal approval). Estimate what percentage of total losses fit the "went quiet right after the deck" pattern.What changes
Free pitches still prove expertise and win business on their merits, but prospects walk away with a reason to hire the agency instead of a blueprint they can hand to an AI tool and execute themselves — and win-rate erosion on high-effort discovery pitches stops being invisible.
A prospect asked for a growth audit before signing. The agency delivered twenty pages: what's broken, why, and what to do about it, broken into clear, numbered recommendations. Two weeks later, the prospect went quiet. Not hostile — just gone.
Three months after that, someone on the account team noticed the prospect's site had quietly implemented four of the twenty recommendations, in roughly the order they were listed in the deck. Nobody signed anything. Nobody stole a file. The prospect's in-house marketer just read the audit, opened Claude, and asked it to turn each recommendation into a build plan.
The audit was always the hard part — that's exactly the problem
Any agency worth hiring can tell a prospect something true and specific about what's broken in their funnel, their content, their site. That diagnostic skill, built on pattern-matching across dozens of accounts, is genuinely hard to fake and genuinely valuable to give away as proof of competence. The trouble isn't that the diagnosis is good. It's that a sufficiently good diagnosis, handed over as a structured document, is now a usable brief — and briefs are exactly what AI tools are best at turning into action.
It doesn't look like theft, so nobody treats it as a risk
Nobody on the agency side frames this as losing work to a competitor, because there's no competitor in the story — just a prospect, a free document, and a tool that used to only summarize things and now builds things. There's no RFP to lose, no pitch to be out-priced on. The deal just goes quiet, and it looks exactly like every other lost pitch, so nobody connects it to the deck.
The fix isn't a worse audit — it's a different kind of leave-behind
Making the free audit vaguer to protect it defeats the point; a vague audit doesn't win pitches either. The fix is separating what proves expertise (the diagnosis, the reasoning, the specific account context) from what constitutes an execution plan (the sequencing, the resourcing, the trade-offs between what to do first versus what to do eventually). The first can stay generous. The second is the engagement, not the pitch.
Delivery format matters as much as content
A narrated walkthrough and a structured leave-behind document can contain the exact same findings and produce very different outcomes, because one is a document an AI tool can parse into a build plan and the other is a conversation a prospect has to actively reconstruct from memory or notes. Proving expertise doesn't require handing over a machine-readable brief.
Bottom line
The free audit isn't broken as a sales tool — it's just operating on an assumption that stopped being true. Diagnosis and execution used to be separated by a skill gap wide enough that giving away the first was safe. AI narrowed that gap enough that the deck itself can be the entire engagement for a prospect willing to do the assembly. The agencies that keep winning with free audits aren't the ones giving away less — they're the ones who noticed which half of the deck was actually doing the selling, and stopped handing over the other half for free.