guava
Your assessment at a glance
7 workflows ranked: 1 Build, 4 Pilot, 2 Don't build. Start with testing guava-stack after each update. Prospect research, post drafts, the Friday digest and posting review already run on skills. Desktop app hours fill in from Screen Time, and the full report is ready around October 28.
- Hours/week on Build + Pilot work
- Populating
- Build
- 1
- Pilot
- 4
- Don't build
- 2
Time spent now on Build and Pilot workflows, from how often and how long each takes. How much AI saves will need to be tested.
What this report draws on
- Your guava-hub and Notion Memory Populating · ready around Oct 28Workflows your assistants pick up as you work. 6 days so far.
- Email (Gmail) ReadyAbout *** emails in 30 days, mostly notices from your tools.
- Calendar Ready** meetings in 30 days, grouped by type.
- Browser history (Brave) Ready17,489 visits in 30 days, totals by site only.
- Screen Time Populating · ready around Oct 27Desktop apps such as Excel and Outlook. Turned on Oct 20.
- AI memories (Codex, Claude Code) Ready49 Codex work sessions in 30 days, counted by topic.
- Your answers SkippedFive quick questions give the hours a report can't measure.
- Live interviews with Guava Not includedWhy work stalls and who waits on whom. Only in the full AI Assessment.
Your tech stack
- X · 19.6 hours of browsing in 30 days
- GitHub · 13.5 hours, website and guava-stack code
- workwithguava.com · 11.5 hours on your own site
- LinkedIn · 10.3 hours
- Marketplace ******** · 4.5 hours, most of it an AI agent working for you
- Gmail · 1.9 hours, about *** emails in 30 days
- Codex and Claude Code · 49 Codex sessions in 30 days
- Notion · guava-hub, memory, source of truth
- Desktop apps · Populating · ready around Oct 27
- Personal sites ******** · ** hours
Your workflows
| Workflow | Verdict | Why | Hours/week spent |
|---|---|---|---|
| Testing guava-stack after each update | Build | Same checks every time, all digital, and a failed check is easy to spot. | unknown |
| Benchmark runs and publishing | Pilot | Runs follow written rules. The one uncertainty is the judgment calls mid-run, such as whether a failed run counts. | unknown |
| Website and search setup | Pilot | Digital and repeatable, but the public sees the result. An assistant drafts, you approve the preview. | unknown |
| Social visuals and video edits | Pilot | Worth trying for banners, covers and rough cuts. The uncertainty is taste: you'll still pick the final. | Populating · ready around Oct 27 |
| Client AI Assessments | Pilot | Your core paid work, and a client sees the result. Start with one part you repeat every time. | unknown |
| Marketplace profile and proposals | Don't build | The overhaul is done and proposals are rare. Revisit once you send them every week. | unknown |
| X replies | Don't build | Relationships, typed for each thread. Keep these by hand. | unknown |
Build = worth automating now
Pilot = try on a small scale first
Don't build = keep doing it by hand for now
Sizes: Skill (one person) · Team rollout · System build (inside your PSA, CRM or ATS) · Service to resell
What I don't know yet
- How long each workflow takes you. Screen Time fills in desktop hours by Oct 27; the rest needs your answers.
- Your priorities: what you most want off your plate first.
- How many paid AI Assessments you run a month (********).
Next steps
- Answer the five quick questions next time, or now. They're saved, so you only answer once.
- Run this again around October 28 for the full report.
- Build the guava-stack update check as a skill.
What this report can't see
Your apps, inbox and history show where the time goes. They can't show why work stalls, who waits on whom, or what lives only in someone's head.
- Client AI Assessments are your core paid work. Where does one stall: booking interviews, waiting on client data, or writing it up?
- Benchmark runs: 4 sessions in 30 days. Which steps do you redo by hand after a run, and what makes you rerun one?
- LinkedIn, X and ******** take hours every week. Which conversations turn into booked calls, and what happens in between?
- What do you redo because an AI tool got it almost right?
Guava's full AI Assessment answers these through interviews with you and the people who do the work: $4,000, two weeks, credited in full toward the next project.