Share
Explore

icon picker
Product‑Page Content Revamp – Marketing Input Workbook

IMG_0134.HEIC

Version 0.2 – 3 Jun 2025

1. Why you’re seeing this

We are building an automated pipeline that turns our existing (sometimes outdated) product‑page content into copy and images that are:
medically accurate
regulation‑compliant
on‑brand and conversion‑optimised
The automation needs structured inputs from Marketing before it can run. Everything below is written in plain language and shows an example next to every field you need to fill. If a row is blank, please complete it once for each product and return the file. (Feel free to duplicate the tables in a Google Sheet.)

2. One‑Page Master Checklist ✅

Table 1
#
What we need (one row per product)
Example entry
1
Product page URL / ASIN
2
Official product name
Kairali Ashwagandha Capsules
3
Ingredient list (see Table A)
—see Table A–
4
Target consumer segment (Table B)
—see Table B–
5
Top 3 competitor links & key claim
Himalaya Ashwagandha – “Reduces stress”…
6
Brand‑voice & style‑guide link
7
3–5 SEO keywords
ashwagandha benefits, stress relief capsules
8
Any claims already approved by Legal
“Helps manage day‑to‑day stress”
There are no rows in this table
💡 Tip: If several products share the same brand‑voice link, just point to the same URL each time.

3. Tables to copy‑paste

Table A – Ingredient list

Table 2
Ingredient
per serving (mg)
Grade / Extract ratio
Source (Organic/Conventional)
Ashwagandha (Withania somnifera)
500
5:1 KSM‑66® extract
Organic ‑ Bangalore
There are no rows in this table

Table B – Consumer segment / persona

{
"segmentName": "Millennial Women 25‑35",
"demographics": "Female • 25–35 • Tier‑1 cities",
"needState": "High stress + poor sleep",
"shoppingDrivers": "Natural, non‑habit forming"
}

4. How your inputs feed the agents

Table 3
Agent in the pipeline
Fields Marketing must supply
Why it matters
Ayurvedic Science Expert
Ingredient list (Table A), Product name
To map classical Ayurvedic texts to each ingredient
Modern Science Expert
Ingredient list
To pull modern clinical studies
Marketing Positioning Agent
Consumer segment (Table B), Top 3 competitor claims, Brand‑voice link
Crafts USP / Benefit / RTB grid that feels on‑brand and differentiates us
Marketing Copy Agent
SEO keywords, Tone‑of‑voice note (1‑2 sentences)
Generates title, bullets & long copy optimised for search & conversions
Graphic‑Design Brief Agent
Visual style reference (mood‑board link or brand‑guide page #)
Writes designer‑friendly briefs for hero image & infographics
There are no rows in this table

5. Ready‑to‑use Templates (copy the JSON / tables)

We’ve pre‑built machine‑readable templates so you don’t have to guess the format.
Ingredient CSV – just grab the heading row from Table A.
Persona JSON – copy Table B.
SEO keywords – comma‑separated plain text.
If you prefer Google Sheets, keep one tab per product and use the same column names; the tech team will convert it automatically.

6. Visual Agent Flow (schema diagram)

Below is a high‑level flowchart showing how the agents hand off to each other. (If you’re viewing this file in a Mermaid‑enabled tool such as GitHub, the diagram will render automatically; otherwise, follow the text arrows.)
graph TD
subgraph Inputs
A1[Existing Content] --> A
A2[Ingredient List] --> A
A3[Consumer Segment] --> A
A4[Competitor Data] --> A
end

A(Input Aggregation) --> B[Ayurvedic Science Expert]
A --> C[Modern Science Expert]

B --> D[Marketing Positioning Agent]
C --> D

D --> E[Marketing Copy Agent]
D --> F[Graphic‑Design Brief Agent]

E --> G[Medical Compliance Agent]
F --> G

E --> H[Regulatory Compliance Agent]
F --> H

%% Iteration loops – dashed lines show change‑requests
G --|Change Req|--> E
G --|Change Req|--> F
H --|Change Req|--> E
H --|Change Req|--> F

G --> I[Output Packaging]
H --> I

7. Detailed Input Reference (fill one row per product)

This expands the quick checklist into explicit fields the automation expects. Copy the headings into a Google Sheet or CSV.
Table 7
Field name (exact)
Description
Example value
Consumed by agent(s)
product_name
Official marketing name (≤70 char)
Kairali Ashwagandha Capsules
All
asin_or_url
Amazon ASIN or full PDP URL
B09XYZ
Input aggregator
ingredient_csv
Link (or G‑Sheet ID) to Ingredient table (see §3 Table A)
Ayurvedic, Modern
consumer_segment_json
Link or inline JSON blob of persona (see §3 Table B)
{…}
Positioning
competitor_links
1–3 URLs separated by ;
https://…; https://…
Positioning
competitor_claims
Matching list of short claims
“Reduces stress”;…
Positioning
brand_voice_url
PDF / G‑Doc of tone‑of‑voice guide
Copy, Design
visual_style_ref
Mood‑board or style‑guide page #
Slide 12 in BrandGuide v2
Design
seo_keywords
Comma‑separated list
ashwagandha benefits, stress relief capsules
Copy
preapproved_claims
Any claims vetted by Legal/Reg
“Helps manage stress”
Compliance Agents
regulatory_class
FSSAI category or AYUSH type
“Nutra – Ayurvedic Proprietary”
Regulatory
batch_size
# of SKUs to process together (int)
1
Orchestration
There are no rows in this table
Tip: You can add additional columns; the pipeline ignores unrecognised headings.

8. SOP Cheat‑Sheet for Each Agent (for visibility)

Table 8
Agent
What it does (internal steps)
Success criteria
Auto error handling
Ayurvedic Science Expert
1. Parse ingredient_csv.2. Query Classical texts DB.3. Return mechanisms & citations in JSON.
JSON schema validates.
On parse error → send error_ing_list flag and halt product.
Modern Science Expert
1. Resolve ingredient synonyms.2. Look up PubMed & AYUSH research API.3. Score evidence, output top 5 studies/ingredient.
≥1 study per active ingredient.
If <1 study, mark low_evidence but continue.
Marketing Positioning Agent
1. Merge Ayurvedic+Modern outputs.2. Map consumer need‑states.3. Fill USP/Benefit/RTB YAML template.
All YAML keys present.
On YAML error → retry 2×, then escalate.
Marketing Copy Agent
1. Read positioning YAML.2. Generate copy pack v1.3. Run style‑guide linter.
Linter score ≥80%.
If <80%, auto‑revise tone up to 2×.
Graphic‑Design Brief Agent
1. Parse positioning YAML + copy.2. Produce JSON brief with 3+ image slots.3. Validate against style refs.
JSON schema passes.
Missing style ref → default to brand palette.
Medical Compliance Agent
1. Run medical checklist (# of disease claims, dosage accuracy).2. Flag issues.
0 critical fails.
Auto‑edit minor dosage typos; else return change‑request.
Regulatory Compliance Agent
1. Check FSSAI/AYUSH rules per regulatory_class.2. Validate disclaimers, font sizes.
Pass checklist.
Auto‑append mandatory disclaimer if absent; else change‑request.
Output Packaging
1. Collect final CopyPack + design briefs.2. Export ZIP bundle.
ZIP checksum >0 bytes.
Retry once on packaging error.
There are no rows in this table
SOPs are coded into the pipeline – Marketing just needs to supply the inputs above.

9. Glossary (plain English)

USP – the single most compelling reason a customer should pick us over others.
RTB (Reason to Believe) – evidence that backs up the USP (e.g., clinical study, Ayurvedic citation).
Tone of voice – how the brand “sounds” (e.g., serious medical, warm & friendly).

10. Iteration Loop & “Edge‑vs‑Compliance” Guardrails

When either Medical (G) or Regulatory (H) Compliance issues a Change Req flag, the system automatically routes their structured feedback back to the Marketing Copy (E) and Design Brief (F) agents. The loop is governed by these rules:
Table 4
Column 1
Column 2
Column 3
Parameter
Value
Rationale
Max automated loops
3
Prevents infinite ping‑pong; 4th fail triggers human triage.
Compliance pass threshold
100 % critical, ≥95 % minor
Must obey law and avoid red‑flag medical claims.
“Brand Punch” score
≥75 % (internal NLP metric)
Keeps wording vivid; if score drops below 75 %, system nudges copy back toward original tone before next compliance check.
Allowed claim removals
Up to 50 % of benefit bullets
Ensures we don’t strip all selling points; encourages re‑phrasing not deletion.
Escalation path
Slack channel #content‑triage + Jira ticket
Puts Medical, Regulatory & Marketing in one thread for manual resolution.
There are no rows in this table

How the loop works

Change‑Req payload lists offending text snippets + rule ID (e.g., FSSAI‑4.5 or DiseaseClaim‑Critical).
Copy/Design agents re‑draft only the flagged sections, preserving brand voice & SEO keywords.
NLP engine recalculates Brand Punch; if <75 %, copy is enriched (tone verbs, emotional language) before resubmission.
On the 3rd failure, the pipeline pauses and notifies humans to decide whether to soften claims or provide more evidence.
Want to print your doc?
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
CtrlP
) instead.