How to Turn One Verified Customer Result Into a Case Study, a LinkedIn Post and a Newsletter in the Founder's Own Voice

A good customer result travels badly. The case study says 45% of tickets were resolved without a human in 60 days. The post written a week later says almost half. The newsletter drops the 60 days. A sales deck rounds it to 50%. Nobody lied, yet the customer who approved one figure now sees three, and the buyer who checks finds that they do not match.

The second problem is voice. Content drafted from a case study tends to read like a case study: a label, a result, a hashtag, a closing question anyone could have written. Senior readers can tell when nobody in particular wrote a post. A founder's post earns attention because it says what the founder saw or thinks, and that is the one thing a drafting tool cannot invent without making it up.

So the work has two jobs. One is keeping the facts identical across every asset, with the customer's approval behind each. The other is putting the founder back into the text. The workflow below gives each job to the tool built for it, and leaves the judgement with you.

Three tools from the CRAFT Content connector and one skill from the B2B Personal Branding plugin, in the order shown in the table alongside. The connector runs free in the browser with no sign-in, or in Claude and ChatGPT through its MCP address; the plugin installs free from GitHub into Claude.

The order follows what each tool needs. The case study generator comes first because it asks for the before, the after and the period of every figure, and those answers are what the customer signs off. The repurposer comes second because it builds each version from whole sentences of your source, so the source has to carry the approved figures already. The newsletter builder adds no statistic or customer story of its own, so it can only repeat figures you hand it. The voice audit goes last, in two passes: the post goes in as the repurposer wrote it, the skill marks where the founder's own lines belong, the founder writes them, and the skill reviews that edit.

My own working rule, not the tools': before the first tool, lock the result as full sentences with the figure first, here "45% of tickets were resolved without a human in 60 days. First response time went from 6 hours to 2 minutes.", and paste exactly that into every input. Set the same sector in all three CRAFT Content tools too.

Step 1. Build the case study with the locked sentence and one sector setting

Input to the case study generator in full mode: customer Example Food Delivery Co, food delivery; product Answerloop; the challenge, "a ticket backlog growing faster than the support team could hire"; the solution, AI agents that resolve refund and order-status tickets inside the help desk, with human approval for refunds; the locked result sentence; no quote; and the sector set to SaaS, the setting I then used in all three CRAFT Content tools.

What came back: the first locked sentence, word for word, as the headline, and the line "Sector: set by your_sector as SaaS, customer service software". The situation paragraph judged the team on "first response time and resolution time", and the results section asked for "the value before, the value after and the period it was measured over" and named the measures the results do not yet show: "resolution time, customer satisfaction, backlog age and cost per resolution". It did not write a quote; it listed one item to add, a line from the customer "approved in writing".

How to read it: the sector setting decides whose measures the story uses. The field asks for your own industry, and Answerloop could be filed as AI native or as SaaS for customer service software, so pick the one your buyers would use. With AI native, the situation line judged a food delivery support team on "accuracy on an evaluation set". With the sector left out, the tool read AI native from my inputs. SaaS read as customer service software and judged the team on "first response time and resolution time", the measures a support leader already reports. Keep the figure first in the locked sentence: when my first try began "In 60 days,", the tool split the result at the comma and the headline became "In 60 days".

The period for first response time is still open, so ask the customer before either figure leaves the case study. For the missing quote, use its interview question: "Which channels do customers use to reach support today, and where do tickets pile up?" My own working rule: write the answer down word for word and send it back for approval as the quote. Hand-off: the same locked sentence goes into the repurposer's source.

Step 2. Turn the story into a post and a newsletter section

Input to the content repurposer: a short source in the founder's words, set as a case study, with the locked sentence on a line starting "Results:", the key message "Measure tickets resolved, not tickets deflected", a professional voice, the SaaS setting, and three formats: a LinkedIn post, a newsletter section and quote cards.

What came back: the post and the newsletter section both carry the locked sentence after the label "Result:". The post puts each source sentence on its own line, with both result sentences together after "Result:", and ends "What is your view?" and "#SaaS". The newsletter section is the key message as a heading, one source sentence, three sentences as bullets, the results line and "Read the full case study". The quote cards are five sentences from the source.

How to read it: the "Results:" line matters. In an earlier run with the figures inside ordinary sentences, the tool reported "Not given: a proof line", a proof count of 0, and 2 unused sentences, "Most support automation only drafts replies." and "The tickets that cost the most", the second cut at its colon. The repurposer keeps your sentences "as you wrote them", so the source decides the quality. Two checks every time. The post dropped the last source sentence, about keeping a person on anything that moves money; it survives only on quote card 5. And the quote cards are your own sentences, so never put a customer's name under one.

In this tool the SaaS setting read plain SaaS, not customer service software: its audience notes talk about time to value and renewal rate, and the tag became #SaaS. Hand-off: the post goes to the voice audit as it came back.

Step 3. Build the newsletter issue around the same sentence

Input to the newsletter builder: topic "AI agents in customer support", call to action "register for our support automation webinar", practitioners, educational, conversational, product Answerloop, the SaaS setting, and three key points: measure tickets resolved, not deflected; what Example Food Delivery Co put in place; and the locked sentence.

What came back: four sections (how it works, what the evidence shows, where readers push back, what to check in your own operation), the button, three A/B subject lines and two hooks. The evidence section is the locked sentence, word for word. The pushback is "We already have a tool for this", settled by comparing daily work "using one real team and one real week". The notes flag "Figures with no source label".

How to read it: add one source line, such as the customer's own help desk data and the date it approved the figures. Then rewrite the preview, "For teams in SaaS, the numbers to watch are time to value and adoption by the target team", and the line "Teams in SaaS would judge it on time to value". Your readers run support, not software companies. In my view they report first response time, time to resolution, CSAT, backlog age and reopen rate; use the ones the customer names. The subject line is the topic itself, so test an alternative.

Hand-off: all four assets now carry the same result, as the table alongside shows. In an earlier run I had reworded the result in my key point, and the issue said "cut first response time from" while the case study said "down from". Locking the sentence before the first tool removed that.

Step 4. Audit the post with Handcrafted voice audit, in two passes

Input to the skill: the LinkedIn post exactly as the repurposer returned it, with the request to audit and fix the founder's post for support leaders at consumer apps.

What came back on the first pass: it set the band from its reading levels table, the row for functional leaders reading a social post, upper secondary with a script grade under 13 as a ceiling. Its scan script printed "banned flags 0 | review flags 0" and "no flags", at an approximate grade of 7.0. Its reading pass, which the skill requires because the script does not catch every pattern, found two problems: no sentence says what the writer saw, did or thinks, the "description with no point of view" pattern, and the "Result:" label in place of a sentence opening.

It deleted the label and left the two sentences word for word, so the post now carries the locked sentence exactly. For the missing point of view it wrote a placeholder, not a line in the founder's voice: "[User to add: one or two lines in your own words, from something you saw or did, on why you stopped counting deflected tickets]". The other seven lines stayed as they were. Its questions: the period for the second figure, whether "Most support automation only drafts replies" is opinion or a claim to check, whether #SaaS is a tag your readers use, and whether Example Food Delivery Co approved being named.

Second pass, in the skill's review mode, with an example line written for this guide, not a real founder's words: "For a year I reported deflected tickets to our board. Then a customer's support lead showed me the same people writing in again the next day." The skill did not rewrite. The scan was clean at grade 6.8, the line cleared the point of view flag, and it asked two new questions: did the scene happen as written, and was that support lead from Example Food Delivery Co? A reader will assume so. How to read it: a clean scan is not a clean post. The scan passed a draft with no founder in it, and the reading pass caught it.

1. Pick one customer result you can back with the customer's own data, and write it as one sentence with before, after and period.
2. Run the case study generator in full mode, answer its questions with the customer, and get the story and figures approved in writing.
3. Write a short source in your own words with the locked sentence on a "Results:" line, and run the repurposer with the same sector setting.
4. Build the newsletter issue with the locked sentence as a key point, and add the source line the tool asks for.
5. Run Handcrafted voice audit on the post, write your own lines where it leaves the placeholder, and run it again in review mode.
6. Before anything goes out, compare the result sentence in all four assets, word for word.

1. Rounding the result differently in each asset. Lock one sentence and paste it everywhere.
2. Publishing a figure without its period. The case study generator asks for before, after and period for a reason.
3. Putting words in the customer's mouth. Leave the quote out until the customer gives one and approves it.
4. Posting the repurposer's draft as it came back. It is your sentences rearranged; it has no point of view of its own.
5. Trusting a clean scan. The voice audit's reading pass found the missing founder after the script reported no flags.

• Auditing a B2B Page for AI Answer Engines: FITq, RACE and a BLUF Rewrite, Run on Our Own Page
• How to Roll One Positioning Hub Out to Every Buyer Role and Channel in a B2B Rebrand, With the Hub-Spoke Methodology
• From the Buyer Who Champions You to a Message Sales Will Use: Positioning With IMPACT, CRAFT Content and Revenue Enablement
• All 15 guides