Delivery audit discovery call
Use this skill when you have 30 minutes with the owner of a professional service firm or agency (an existing customer, a closed-lost contact, or a net new prospect) to find the one client-delivery problem worth building a custom solution or pilot for. Produces a quantified pain (annual dollar figure plus opportunity cost in the owner's own words), a price framing delivered on the call, and a booked next step. Also use for "delivery audit", "AI audit call", "discovery call for a custom build", "scope a pilot", "find the gnarliest workflow", "price on the call".
Applies the moment a 30-minute "delivery audit" or discovery call starts with a firm owner. Produces one quantified pain, a price the owner has already justified to themselves, and a next meeting on the calendar before they leave.
The play
Timings are for a 30-minute call. Open references/call-script.md for the exact lines in each block and the net-new variant.
- Open empty (0-2 min). Ask what was on their mind when they accepted the invite. Then stop talking for three full seconds. If they deflect, ask what comes up when you say "client delivery". Do not show a list of categories first; a list boxes them in and you miss the biggest pain.
- Net new or closed-lost: add a 3-minute scene-set (what the firm does, delivery team size, active engagements). You need a named pain by minute 10.
- Existing customer with usage data: bring the data in after they name a pain, as ammunition. It either backs up what they said or surfaces something they forgot. Leading with it boxes them in.
- Anchor the activity's value (2-5 min). Before digging in, get a yes to "this is a critical part of what clients pay you for" and "this is a competitive advantage piece". Without the yes, the later number is just time saved and the price looks expensive.
- Vertical drill, one pain only (5-15 min). Walk the process start to finish, as nitpicky as they can. Then: which step eats the most time, what share of the work, how often, who does it, their loaded cost, what they already tried and why it failed. Collect the numbers; no math out loud yet. A second pain gets parked. Deep on one, not shallow on five.
- Magic wand (15-22 min). "If the [biggest step] were just done for you, what does your day look like?" Then convert: what a new client is worth per month, what a hire would unlock. If they answer small, agree it is huge and ask whether it means more work or headspace for something else. Leave the block with an annual dollar figure for the pain, an opportunity-cost number, and their words on why it matters. Open references/pain-math.md for the capture sheet and the conversion.
- Price on the call (22-27 min). Wait for them to ask. If they have not by minute 24, offer to share what the investment usually looks like. Restate their figures, anchor at the standard range, offer the early-partner rate for this scope, ask whether that feels like the right ballpark. Describe the build in one breath. Open references/pricing-frame.md for the lines and the objection handling.
- Lock and ask for the referral (27-30 min). Yes-ish: one-page scope by a named time, 30 minutes booked this week to walk it through. Maybe: send the summary with the rough numbers and let them sit with it. Always close with the referral ask: another firm owner they respect wrestling with similar stuff.
What good looks like
- First thing I check that most people skip: whether the owner said the value of the activity out loud before I started counting hours. Without that anchor, every number after it is "time saved" and the price conversation becomes a cost conversation.
- Second tell: whether the dollar figure came from their mouth or mine. If I said it first, the proposal has no teeth.
- The mediocre version: a horizontal tour of the whole delivery process, a little digging on five things, no number on any of them, "let me send you some pricing" by email, and "let me know" at the end. It feels like a good conversation and nothing happens.
- I know the output is right when, by minute 22, I can fill in one sentence without guessing: "[Owner] is losing about [annual figure] a year on [one step], which is stopping them from [opportunity], and they said it matters because [their words]."
- I know the call landed when the next meeting is booked before the call ends and the owner treated the price as a ballpark question, not a shock.
- Red flag: the owner is excited about three problems and you leave with none quantified. Go back and pick one.
Rules
- MUST get a named pain before the drill. On a net-new call, by minute 10.
- MUST anchor the activity's value before digging into the pain.
- MUST stay on one pain. Park every second pain for the follow-up.
- MUST leave the magic-wand block with an annual dollar figure, an opportunity-cost number, and the owner's own words on why it matters.
- MUST discuss pricing on the call. Pricing by email doubles the stall rate.
- MUST book the next step before hanging up. "Let me know" is not a next step.
- MUST ask for the referral on every call, as a separate last thing.
- NEVER lead with a category list, a dashboard or an audit report before the owner has named their own pain.
- NEVER do the math out loud during the drill.
- NEVER quote the price before restating their figures.
references/call-script.md
Delivery audit call script
30 minutes. The owner talks; you ask and count. Bracketed text is what you fill in from their answers.
Block timings
| Block | Existing customer | Net new or closed-lost |
|---|---|---|
| Open empty | 0-2 | 0-3 |
| Set the scene | skip | 3-6 |
| Find the pain | inside the open | 6-10 |
| Anchor the activity's value | 2-5 | 10-13 |
| Vertical drill (one pain) | 5-15 | 13-22 |
| Magic wand | 15-22 | 22-26 |
| Pricing | 22-27 | 26-29 |
| Lock + referral | 27-30 | 29-30 |
The net-new call has no usage data to lean on, so the scene-set matters more. Keep it tight: 3 minutes maximum.
Open empty
Existing customer:
- "What were your thoughts coming into this call? What was on your mind when you saw 'delivery audit' on the invite?"
- (If they deflect) "When I say client delivery, what comes up for you?"
- (If still nothing) "What's the part of delivery you wonder if AI should be able to do already?"
Closed-lost or past conversation:
- "Last time we talked it was about [the old topic]. We've since shifted toward [your current focus] for firms like yours. Curious what was on your mind agreeing to this call?"
- "When I say challenges or bottlenecks in client delivery, what's the first thing that comes to mind?"
Net new, after one sentence of context on what you do and why you are talking to firm owners about delivery:
- "So I want to keep this open. When I say client delivery, what comes up for you?"
Then shut up. Three full seconds. Silence does the work.
Set the scene (net new only)
- "Just so I have context, walk me through what your firm actually does. What's a typical engagement look like?"
- "How big is the team doing the delivery work?"
- "How many active engagements right now?"
Listening for: firm size, delivery model, where the leverage might be. Move fast.
Find the pain (net new only)
- "When you think about the work your team actually does day to day, what's the part you wonder if AI should be able to do already?"
- (If vague) "What's the part of delivery that keeps you up at night right now?"
- (If still vague) "Tell me about the last project where something broke or took way longer than it should have."
You need them to name a specific pain by minute 10. Do not move on without it.
Anchor the activity's value
- "Just to make sure I'm tracking, [the activity] is a pretty critical part of your delivery, right? This is what clients are actually paying for?"
- "So this is a competitive advantage piece for you?"
Get the yes before moving on.
Vertical drill (one pain only)
- "Walk me through the process. As nitpicky as you can. You sit down to do [the thing]. Walk me through start to finish."
- "Of all those steps, which one eats the most time?"
- "Like is that 80% of the work? 50%? Just rough."
- "How often does this come up? Per client? Per week?"
- "Who's doing it? You? Someone else?"
- "What's their loaded cost roughly?"
- "Have you tried to solve this already? What did you do?"
- "Why didn't that work?"
Collect numbers. Do not do the math out loud yet.
Magic wand
- "Let's say this was completely taken care of. The [biggest step] is just done for you. What does your day look like now?"
- "What does that mean in actual numbers? What's a new client worth to you per month?"
- "If you could hire someone, what does that unlock revenue-wise?"
- (If they go small) "That's huge. But would that translate to taking on more work? Or freeing up headspace to fix something else?"
Leave this block with: an annual dollar figure for the pain, an opportunity-cost number, and their words on why it matters.
Pricing
See pricing-frame.md. Wait for them to ask; prompt at minute 24 (existing) or 27 (net new) if they have not.
Lock + referral
If yes-ish:
- "I'll send a one-page scope by [tomorrow morning] with the build, pricing, and timeline. Can we put 30 minutes on the calendar later this week to walk through it?"
If maybe:
- "I'll send a summary including the rough numbers. Sit with it. If it makes sense, we scope it properly."
Always:
- "Last thing, totally separate. If anyone comes to mind, another firm owner you respect wrestling with similar stuff, I'd love an intro. Trying to talk to as many of these folks as possible right now."
Don't forget
- One pain only. Go vertical, not horizontal.
- Anchor the value before digging into the pain.
- The magic wand question is the unlock.
- Price on the call. Never by email.
- Lock the next step before they leave.
references/pain-math.md
Pain math: what to capture and how to convert it
Collect during the vertical drill. Convert during the magic wand block. Never convert out loud during the drill; it turns an interview into a negotiation.
Capture sheet
| Field | Question that gets it | Example answer |
|---|---|---|
| The activity | "Walk me through the process start to finish." | Monthly client reporting |
| The biggest step | "Which step eats the most time?" | Pulling data from four sources into the deck |
| Share of the work | "Is that 80%? 50%? Rough." | About 70% |
| Frequency | "How often? Per client? Per week?" | Once a month per client, 12 clients |
| Hours per occurrence | Usually volunteered with the step; ask if not | 3 hours each |
| Who does it | "Who's doing it?" | An account manager |
| Loaded cost | "What's their loaded cost roughly?" | $90 an hour |
| Already tried | "Have you tried to solve this? What did you do?" | A reporting tool, abandoned after two months |
| Why it failed | "Why didn't that work?" | Still needed manual cleanup every time |
The two numbers you need
Annual cost of the pain
hours per occurrence × occurrences per year × loaded hourly cost
Example: 3 hours × (12 clients × 12 months) × $90 = $38,880 a year.
Use the owner's figures even when they look rough. The number only has to be theirs.
Opportunity cost
What they said they would do with the time, in dollars. Prompt with "what's a new client worth to you per month" or "what would a hire unlock". A new client at $4,000 a month is $48,000 a year of upside they are not pursuing.
If they go small on the opportunity, agree that it is huge and ask whether it translates to more work or to headspace for something else. Either answer gives you a second number.
The sentence you should be able to say by minute 22
"[Owner] is losing about [annual cost] a year on [the biggest step], which is stopping them from [opportunity], and they said it matters because [their words]."
If any bracket is empty, go back to the block that fills it. Do not move to pricing without all four.
Common misses
- Counting the whole activity instead of the biggest step. The build will target the step; price against the step.
- Taking "it's annoying" as the value. Annoying is not a number. Ask what it costs or what it blocks.
- Accepting the owner's own time at zero. Ask what their hour is worth to a client, or what they bill.
- Letting a second pain into the sheet. Write it in the margin for the follow-up and go back to the first one.
references/pricing-frame.md
Pricing frame and objection handling
Pricing happens on the call. Stall rate roughly doubles when pricing goes by email, because the owner does the value math alone and without the anchor.
When to open it
Wait for the owner to ask. If they have not asked by minute 24 (existing customer) or minute 27 (net new):
- "Want me to share what the investment usually looks like?"
The frame, in order
- Restate their figures. "Based on the value you just described, [their pain] is costing you about [their figure] a year, plus the opportunity cost of [what they'd do]."
- Anchor at the standard range. "Most customers in similar situations land between [standard low] and [standard high] a month."
- Offer the early-partner rate for this scope. "That said, we're early on this offer and being deliberate about who we work with. Looking for co-partners more than customers. So for you, based on this scope, I'd propose we start at [partner rate] a month. As scope expands, we revisit."
- Check the ballpark. "Does that feel like the right ballpark for the value we just talked about?"
- Describe the build in one breath. "What we'd do is embed an engineer with you for the next few weeks to build a custom solution for [the pain]. Build is included. After deployment, we keep it running, maintain it, improve it. First [scoping period] are no-charge while we onboard and scope."
Then stop talking.
The figures the script shipped with, for a custom AI-build offer to service firms: standard range $5K to $7K a month, early-partner rate $3K to $4K a month, scoping period 10 days. Substitute your own; keep the structure. The anchor must be the standard range, and the partner rate must be tied to the scope they just described, so the discount reads as selectivity rather than weakness.
A lower-ticket pilot variant for an existing customer: a fixed one-month pilot fee, refunded in full if the agreed hours are not saved, rolling into an annual license if they are. State the hours target out loud; it comes from the pain math.
Handling the "is this worth it" pause
Go back to their number, not yours:
- "You mentioned this was costing you roughly [their figure] a year. If this solves it, even partially, does the math work for you?"
This is not pressure. It is the calculation they are already doing in their head, done out loud.
Objections
| They say | You say |
|---|---|
| "Price is too high" | "Fair. What were you expecting, and what would make it a no-brainer?" |
| "Need to check with my partner" | "Totally get it. Would it help if I did a quick 15-minute walkthrough with them? Way more effective than summarizing secondhand." |
| "Need to think about it" | "Totally fair. What specifically are you weighing?" Then send the recap tonight and book 15 minutes to either kick off or close the loop. |
| "We've tried tools for this" | "What did you try, and why didn't it stick?" Usually the answer is the manual cleanup step, which is the step you are proposing to build for. |
| Silence after the price | Ask the ballpark question again. Silence is not a no. |
Isolating price
If the only objection is price, say so and test it:
- "All things remaining the same, if price wasn't a question, do you have any other hesitations?"
If no: "Cool. What would make sense?" You want a value conversation where it is a no-brainer for them; if it is not, it does not make sense for either side.
If yes: that hesitation is the real objection. Handle it first; price was a proxy.