From Italian Memo to Investor Teaser: 5 Steps That Cut Document Prep from Days to Minutes
The 5-phase process that eliminated translation bottlenecks and enforced institutional standards—without inventing a single data point.
Welcome to a new issue of the Unlocking Real Estate Value newsletter. Each week I will provide you with exclusive advice and professional insights to help you realise long-term value through real estate development.
Last week I needed to convert a low-quality Italian feasibility study into an investment teaser for UK institutional investors. The source document was 15 pages of architectural analysis. No financial metrics. No asking price. Just planning status and building specs.
The traditional approach would take 3-5 days. Manual translation. Reformatting. Finding comparable teaser examples. Matching the house style. Chasing missing data.
I completed it in 45 minutes using a structured AI workflow. Here’s exactly how.
Step 1: Set the Non-Negotiable Rules
The biggest risk with AI document work is hallucination. The model fills gaps with plausible-sounding nonsense. You don’t notice until a client questions a figure that never existed.
I created a custom instruction prompt with one core rule: zero hallucination. If data isn’t evidenced in the source document, use a placeholder. Never guess. Never “clean up” inconsistent numbers. Never paraphrase away legal constraints.
Every claim had to include a source reference: page number, section title, table location. If OCR confidence was low, flag it explicitly.
This rule eliminated the biggest AI risk in 30 seconds of setup.
(Full prompt included at the end of this article)
Step 2: Extract and Map Everything
I uploaded three documents:
The Italian feasibility study (15 pages, scanned PDF)
Two institutional teaser examples (to extract house style)
Target audience specs (asking price, investor type, target market)
The AI performed OCR extraction and built a complete source map. Every heading. Every table. Every metric. Page-by-page quality log noting sections with low confidence.
Key numeric fields catalogued: areas (m², mc), planning contributions, building specifications, permit status.
Result: A searchable reference showing exactly what data existed and where it came from. No guessing required.
Step 3: Reverse-Engineer the House Style
Generic teaser templates look generic. Institutional investors recognize house styles. Each advisory firm has formatting conventions that signal professionalism.
The AI analyzed the example teasers and extracted:
Structure: 7-page flow (Cover → Overview → Location → Building → Planning → Transaction → Back cover)
Formatting: Bullet-driven layout, “Key Numbers” table with icons, specific color accents
Tone markers: UK English, sentence case, data-driven language, understated confidence
Conventions: How to present metrics, table formats, disclaimer placement
It produced a “House Style Spec”—10-15 bullets defining voice, structure, and visual standards.
The difference: Style inheritance from real examples, not generic AI assumptions.
Step 4: Get Outline Approval Before Drafting
The AI produced a proposed teaser outline showing:
What content existed in the source (with page references)
What data was missing (flagged with “[TBC / Not evidenced]” placeholders)
7 clarifying questions for genuine gaps (asking price, target returns, operator strategy, transaction structure)
I reviewed the outline. Provided the missing commercial details and brand guidelines.
Critical point: No full draft until the structure was validated. This prevented wasted effort and ensured alignment.
Step 5: Execute with Brand Compliance
The AI generated the PowerPoint teaser applying:
Brand color scheme throughout every slide
House style from the examples (bullet layouts, metrics tables)
Source data from the Italian memo with page references maintained
“[TBC]” placeholders for missing financial metrics
It converted slides to images for visual QA. Identified text wrap issues and icon contrast problems. Fixed them. Rendered cleanly across all 7 slides.
Final deliverables:
Investor Teaser (PPTX)—institutional-grade presentation with branded design
Analysis & Source Map (MD)—complete OCR log, house style spec, source references
Placeholder table—every “[TBC]” field listed for completion before investor circulation
Time investment: 45-60 minutes from upload to final deliverable.
What This Workflow Eliminates
Without AI: 3-5 days of manual work
Translate 15 pages
Hunt for comparable teaser examples
Guess at formatting conventions
Manually check every data point
Risk introducing errors during reformatting
With this workflow: 45-60 minutes of structured execution
Zero hallucination through placeholder protocol
Style inheritance from real examples
Complete source mapping maintained
Brand compliance enforced
Missing data explicitly flagged
The AI doesn’t replace expertise. It enforces rigor.
Key Takeaways
Here’s what you need to implement this:
Set the zero-hallucination rule first—use placeholders for missing data, never guess
Build a source map—every data point linked to page/section from original document
Extract style from examples—don’t use generic templates, inherit real house styles
Get outline approval—validate structure before drafting to avoid wasted effort
Flag gaps explicitly—maintain a table of “[TBC]” fields requiring completion
Italian deals need international capital. Document quality shouldn’t be the bottleneck. This workflow gives you institutional credibility without inventing a single data point.
That’s all for today.
See you next week.
— Carlo
Founder and Managing Director Benigni
