How Joshua Tiernan had Claude run about 80% of his sponsorship business from a 'Next Touch Date' field

The problem

He says his “business runs on a sponsorship model.” His sales process and sponsorship deliverables “take up a large part of my time.” He did not want to build “a sequence with follow-ups on specific dates.” He says every customer or potential customer “is dealt with in an entirely bespoke way.” That depends on the deal size, the type of company and whether they open the emails. He also says that when he ran everything during work hours, he “was running out of usage.”

The blueprint (as he describes it)

1. Build while doing the work. He did not plan the system in advance. He did the day-to-day work with Claude and turned repeated steps into processes as he went. His advice is to run everything past Claude first and ask it what it would decide. When you disagree, push back “so that a) you learn from your errors, and b) it learns your business.”

2. Use HubSpot for the pipeline, not sequences. HubSpot manages “the sponsorship pipeline and availability of sponsorship assets.” He uses it this way “rather than building sequences to handle follow-ups.”

3. Add a “Next Touch Date” field. He added a custom date field in HubSpot titled “Next Touch Date.” Claude runs through it on a daily ritual. When the next touch date is today, Claude drafts a follow-up email based on rules he set in Claude Skills.

4. Write hold-and-release rules for sponsor dates. Under those rules, Claude “pitches dates with specific deadlines for how long we can hold it for.” It “will also release the dates when we’ve not heard back from them.” He says this “helps maximize the sponsorship slots being sold.”

5. Keep deliverables in Notion with review dates. Notion has four databases:

   • A sponsors database for leads who have signed up.
   • A separate deliverables database for each sponsor’s deliverables.
   • An ideas and strategies log. Ideas from him or Claude are logged there automatically.
   • A tasks database for actions he has to take.

   “I don’t look at any of these databases.” Each item has a review date. When an item’s date is today, Claude decides whether it should be done. It then works out how much time he has and puts his tasks into his to-do list app automatically.

6. Move processes to scheduled tasks. He first triggered processes manually. He then moved them to scheduled tasks so they run at different points and spread his usage. He says Claude “works every evening and every weekend.” Some research tasks run twice a month. Others run every day, or sometimes twice a day.

7. Split tasks into sub-tasks. Prospect outreach involves “research, scoring and then outreach.” He suggests running research as a weekend scheduled task. His example: if you need 5 new prospects per day, you need 25 per week, and ideally more. Scoring and outreach run daily, and Claude drafts the emails.

8. Review each morning and send yourself. The daily ritual runs before he opens his computer. It drafts replies to his inbox, checks deals that need a follow-up, assesses the week’s strategy and gives him the day’s to-do list. He then:

   • sends priority emails, which are already drafted;
   • sends LinkedIn connection requests from a list of links Claude puts in his to-do app;
   • does any other tasks Claude has set him.

   “Nothing goes out without me.”

9. Clean up periodically. He recommends sending Claude this prompt from time to time: “Is our system optimal and do we have any redundant tasks or skills that can be cleaned up? Is there anything we can do to improve the current setup? Are there any gaps that we’re missing?” He says that when Claude gets worse, the usual cause is over-engineering. Overlapping skills and scheduled tasks try to do too much in one pass. His fix is to split the work into more tasks or skills.

Stated results

• None stated in measured terms. The post gives no revenue, deal count or reply numbers.
• His own claims, not measured: Claude “runs about 80% of it.” The Claude Max subscription “costs me £90 per month” and is “doing the work of probably 10 to 20 full-time humans.”
• He says, “Each week, the repetitive tasks become more automated than the previous week.”

Caveats

• This is his own newsletter. The issue carries a paid sponsor slot for ZooWork, an AI agent product.
• The issue promotes his paid Pro Membership, which includes the Tiny Empires Method course at nocodefounders.com. It also includes a sign-up line for “free guides.”
• “About 80%” and “10 to 20 full-time humans” are his estimates, not measurements.
• He says he is “looking at making the entire system shareable.” No kit or template exists in this post.
• The 5 and 25 prospect figures are an illustrative example, not his actual volumes.
• He does not name the to-do list app or the KPI spreadsheet tool. He does not share the Claude Skills rules themselves.
• He says the post “was written by me, not Claude.”
• Handles: X @Joshua_Tiernan (verified 2026-10-24: linked from the Tiny Empires about page; X bio reads “Founder of @nocodefounders | 6 exits”. joshuatiernan.com links a different, near-empty @JoshuaTiernan account, which we do not tag). LinkedIn: joshuatiernan (verified: linked from joshuatiernan.com and the Tiny Empires about page). Brand X: @nocodefounders (verified: linked from nocodefounders.com). No Bluesky, Threads, Instagram, TikTok or Mastodon handles verified.

Sources

• https://tinyempires.substack.com/p/i-made-claude-my-ceo (primary: first-person Tiny Empires issue, published 2026-10-16)
• https://www.joshuatiernan.com/ (his own site: background on No Code Founders and Tiny Empires)

---

Want agents like these working your distribution? See the Agent Stack Teardown. Get new teardowns weekly: subscribe (double opt-in).