How To Use: Tri-Backlog Resolver (Level 2)
This companion web application resolves the unresolved questions (PVP, Customer Persona, Value Proposition, or Brand Identity backlogs) identified during your initial 10-minute Level 1 voice onboarding interview.
-
1. Upload Level 1 File
Drag and drop your saved Level 1 text file (e.g.
Expert_AI_Prompts_Matthew_Bulat_CustomerPersona_17Aug2026.txt) into the upload card below. - 2. Resolve SWOT Decisions Identify personal strategic trade-offs. Fill out Pros, Cons, and SWOT textareas to lock in final decisions.
- 3. Collect Google & Deep Research Copy custom-tailored prompts to Google Gemini, ChatGPT or qwen.ai (up to 3 tabs per provider). Paste raw answers back to generate 3-point tactical summaries.
-
4. Export _Level2 Text
Appends resolved intelligence, strips outstanding blocks, and exports file named with
_Level2.txtsuffix.
Before you start — two ways to work faster
Tip Work the modules out of order — start the Deep Research first.
Deep Research takes 5–20 minutes per question. If you run the page top to bottom, you finish Modules 1 and 2 and then sit and wait. Reverse it:
- Upload the Level 1 file and scroll straight to Module 3. Copy each Deep Research prompt into its own AI tab and start them all.
- While they run, work Module 1 live with your client — the Personal Decisions, Pros, Cons and SWOT.
- Then Module 2, the quick Google lookups.
- Come back to Module 3, paste the finished research, and summarise.
A twenty-minute wait becomes twenty minutes of billable interview time.
Advanced Consultants: run all four documents in one sitting — four tabs, one session.
A full Business DNA engagement is four Level 1 files: Personal Value Proposition, Customer Persona, Business Value Proposition and Brand Identity. You do not need four appointments.
- Open this page in four browser tabs and load a different Level 1 file into each one.
- Each tab is completely independent — its own file, its own workspace, its own
_Level2.txtexport. They do not interfere with each other. - Once a file is loaded, each tab renames itself to the client and document type, so you can tell the four apart on the tab bar.
- Ask your client every Personal Decision across all four documents in one sitting, switching tabs as you go.
- Start the Deep Research in all four tabs first (see the Tip above), keeping to a maximum of three research tabs per AI provider — split the rest across Gemini, ChatGPT and Qwen.
Never open two tabs on the same Level 1 file. Each tab holds its answers in its own working memory. If two tabs share a file, only the tab you press Compile in is saved — the other tab’s decisions are lost with no warning. One file, one tab. The same applies to refreshing or closing a tab mid-session.
1. Outstanding Personal Decisions
Resolve trade-offs using structured SWOT & Pros vs. Cons analysis
2. Quick Google Lookups
Batch research fast factual data points via web search assistants
Batch Search Prompt Builder
Use Google Gemini for this stage. Run this prompt there to resolve all fast lookup queries in a single pass. Other assistants append source lists and reference numbers that are captured as findings.
Structured Search Insight Cards
3. Deep Research & Analysis
Conduct extensive, custom B2B industry research and synthesize 3-point tactical summaries
Usage & Limits — read once before you start
Tip Run up to 3 research questions at once per AI provider.
Note In Gemini, select a Pro model first — Deep Research is not offered on Flash-Lite.
Caution Deep Research is the heaviest drain on every provider’s usage limit.
Ready to Compile Level 2?
This will stitch your resolved Decisions, Google Insights, and Deep Research summaries back into your Level 1 document structure and download a pristine _Level2.txt file.
Optional · Client Report Prompt
Turn this session into a presentable Word document your client can take away
The _Level2.txt export above is the machine file that feeds the
Super AI Prompt Creator.
This button builds a prompt that turns the same session into a readable report instead:
a cover page, a two-page session draft of your document,
then one page per decision, one page of Google lookups, one page per research question, and a Business DNA appendix.
The appendix reproduces your Level 1 Business DNA exactly as captured. Any field left blank appears as an em dash, so gaps are visible before the report reaches your client.
Lock in at least one decision, insight or research summary above to enable this.
Paste this into Claude and ask it to run. It will return a downloadable Word file.
The two-page body is a session draft — for the polished, fully-developed document,
run your _Level2.txt export through the Super AI Prompt Creator.
Consistency Check — 3-Stage Review
Upload an already-compiled _Level2.txt file (including one you've hand-edited to apply a resolved decision) and this runs a three-stage review before you use it downstream: 1. Self-QA automatically scans for values that look out of sync — a MACHINE DATA BLOCK field that doesn't match Section B's own copy, an old placeholder that a decision already resolved, or an identity field that never made it into the free-text Narrative. 2. Interim Yes/No puts every finding in front of you as a one-click confirmation — nothing is changed automatically. 3. Final Update applies only what you confirmed, logs the review inside the file itself as a new Section G, and downloads a reviewed copy alongside your original.
Detection is best-effort text pattern matching, not a language model — it's tuned to catch the exact class of miss this was built for (a resolved decision that didn't propagate everywhere), not to guarantee full accuracy. Always read each finding before confirming.
Stage 1 · Self-QA
Stage 3 · Final Update
Applies every confirmed change, logs the review as a new Section G, and downloads a reviewed copy (original file is left untouched).