# IPflair Engagement Review — Quote, NDA, IDF
**Date:** 2026-06-12 · **Skill:** verixa-patent-attorney (meta-orchestrator) · **Matter:** counsel/filing-agent vendor evaluation
**Inputs:** IPflair email quote (Khaja Haruni, Executive Director) · NDA (2 pp) · IDF (generic) · process flowchart · corporate profile · Verixa tracker + FINAL Filing Recommendation 2026-05-29

## A. Matter Intake
- Matter type: filing-agent / counsel engagement review (§7 #11–12, Sub-Workflow H framing) + trade-secret counsel (K, inline)
- Jurisdiction: 🇮🇳 India
- Verixa position: 8 Band-A provisionals **fully drafted** (Form 2 official format, with drawings, repo-verified), FD target 2026-06-15; readiness 72/100 as-is, 82/100 with counsel Claim 1 review (₹15–25K budgeted)
- Deadline pressure: FD in 3 days
- Tier: T2 (engagement + risk review; outside counsel decision rests with founder)

## B. Model Selection Note
- Model used: claude-fable-5 (session) · Effort: high
- Why: vendor/legal-risk review, document analysis, no repo review required
- Output status: **attorney-review-ready** (engagement recommendation), with [Unverified] items flagged
- Stronger model review recommended: no — but outside counsel judgment replaces nothing here

## C. The Decisive Finding — Service Mismatch

**IPflair's quoted provisional product does not match what Verixa needs, and accepting it as-quoted would damage the portfolio.**

Their provisional (₹8,000): *"does not include claims and is filed with an abstract of idea with 150 words only, and a brief description explaining broader scope of your idea in a paragraph."*

That is a **thin/skeletal provisional**. Verixa's provisionals are full Form 2 specifications with drawings, drafted to maximize §11(2) priority ("fairly based on the matter disclosed"). Filing a 150-word abstract instead would collapse priority coverage to almost nothing — the complete spec in 2027 could not claim priority for anything not fairly based on that paragraph. **Verixa must file its own drafted specifications, not IPflair's standard product.**

What Verixa actually needs from a firm:
1. **Review** (not drafting): Claim 1 / spec review vs Semaladhari prior art — the 72→82/100 item
2. **E-filing as agent of record** of already-drafted Form 2 specs (Form 1, Form 26 POA; fee verification at ipindia.gov.in)
3. **2027 complete-spec drafting** — the real engagement (8 families, claims, §3(k)-hardened)

None of these is what the quote prices. **Re-scope before engaging.**

## D. Quote Analysis (per application — Verixa has 8)

| Item | IPflair fee | Read |
|---|---|---|
| Novelty search ₹10K | Optional | Verixa has prior-art search + FTO opinion folders already; likely redundant for Band-A. Skip or sample-test on one family. |
| Provisional draft+file ₹8K | **Do not buy as scoped** | Need filing-only price for pre-drafted specs — ask. Should be materially lower. |
| Complete spec ₹35K | The real cost | ×8 = ₹2.8L + GST. For AI/software CRI claims this price point implies volume practice — quality test before committing 8 families. |
| RFE ₹8K + govt ₹4K | Standard | Per family. |
| Form 3 ₹5K | Note | Rule 12(2) as amended 2024: event-triggered, not periodic — don't pay for unnecessary periodic filings. |
| FER response ₹35K, hearing ₹35K | Standard range | Quality matters most here; §3(k) defense lives or dies in FER/hearing. |
| **Disbursement ₹30,000** | **Query** | Unexplained line item. Ask exactly what it covers. |
| Excess fees note | **Material** | Official fee covers 30 pp / 10 claims. Verixa specs + drawings likely exceed 30 pp now, and complete specs will exceed 10 claims. Get the per-page/per-claim schedule in writing — across 8 families this is a real number. |

**Lifecycle estimate, 8 families through final decision: ~₹13–14L professional fees + 18% GST + official fees + excess-page/claim fees.** [Assumed — arithmetic on their quote ×8; verify.] This must sit in the seed use-of-funds legal/IP line.

**Government fee tier flag:** ₹1,600 assumes Individual/Startup/SME. Today the applicant is Vimal personally (natural person — fine). After assignment to the funded entity (a fundraise closing condition), the entity's fee status must be re-established — startup/small-entity claims need Form 28 and eligibility verification, and a US Inc applicant may not qualify. `Current source required — verify fee category at assignment with counsel.`

**Timeline flag:** their provisional TAT is 10–15 working days; FD is 3 days out. The right question for them: *"Specifications are drafted. Can your agents review Claim 1 of two lead families and e-file all 8 as drafted by Jun 15 — or within a 1-week slip?"* If no, self-file Jun 15 (72/100 path, acceptable per tracker) and engage them only for the 2027 complete-spec stage.

## E. NDA Review (Sub-Workflow K, inline)

One-way NDA, governed by Indian law. Signable after fixes, with discipline:

| Issue | Risk | Action |
|---|---|---|
| **Marking requirement (Clause A)** | Anything not marked "confidential" in writing at disclosure (or confirmed within 1 month for intangible disclosures) is **unprotected** | Standing legend on every document + transmittal log. This is the operative trap. |
| **No term/duration** | Silent on survival; ambiguity in enforcement | Request explicit term: ≥5 years + perpetual for trade secrets |
| **No return/destruction clause** | Materials persist at vendor after engagement | Request return/destruction on termination |
| **No injunctive-relief / venue clause** | Enforcement friction | Request venue (Bengaluru courts fine) + injunctive relief acknowledgment |
| B(1) exclusion lacks written-record proof | "Already knew it" defense is easy | Request "as evidenced by written records" |
| Use restricted to "patent elaboration & application service" | Good — purpose limitation is present | Keep |

**Trade-secret discipline regardless of NDA:** never send the held-out items (Mira prompt templates, production DOMAIN_RELEVANCE_MATRIX values, DQG weights, tuning constants, training-data composition). They are out of the provisionals by design; they stay out of vendor hands too.

## F. IDF — Do Not Fill It

The generic IDF ("upon completion becomes an official document and you are legally bound") is for inventors with an idea and no draft. Verixa has complete specifications. Filling the IDF would (a) create a second, looser disclosure record that can contradict the specs, and (b) over-disclose. **Provide the drafted Form 2 specs themselves, per family, under signed-and-marked NDA — nothing else.**

## G. Firm Credibility ([Unverified] items)
- Registered patent agents listed with IN/PA numbers — verifiable on the IP India agent register before engagement. `Verify.`
- "3000+ filings", Karnataka Govt MOU, IPExcel San Francisco parent — `Unverified — marketing claims.`
- Open competence question: CRI/§3(k) depth for AI/software claims. Ask for: examples of granted CRI/AI patents they prosecuted, FER responses on §3(k) objections, and familiarity with *Lava v Ericsson* (Delhi HC 2024). Their thin-provisional standard product signals volume practice; the complete-spec stage needs deep-tech practice.

## H. Recommended Engagement Structure

**Stage 1 (this week, ₹15–35K):** Sign NDA (with fixes, or as-is + strict marking discipline if time forces it). Paid review of Claim 1 vs Semaladhari on P02 + P03 only. Ask the filing-of-pre-drafted-specs question. This is simultaneously the quality test.
**Stage 2 (FD):** If they pass and can move: e-file all 8 as drafted, IPflair as agent of record. If not: self-file Jun 15; no harm — agents can be appointed later (Form 26).
**Stage 3 (by 2026-12, decision; 2027 execution):** Complete specs. Negotiate a portfolio rate for 8 families; run 1–2 comparison quotes from CRI-specialist firms before committing. Do not award all 8 to an untested firm.

**Kill rule:** if the Stage-1 review work product is shallow (no claim-element mapping against Semaladhari, no §3(k) framing), pay the invoice, take the learnings, and go elsewhere for Stage 3.

## K. Red Flags Summary
1. Thin-provisional product — portfolio-damaging if accepted as scoped (decisive)
2. NDA marking trap + no term/return clauses
3. Generic IDF invites over-disclosure
4. Unexplained ₹30K disbursement
5. Fee-tier change risk at IP assignment to funded entity
6. CRI/AI prosecution depth unproven

## M. Confidence
**78%** (engagement recommendation). Raises: IP India agent-register verification; their answer on filing pre-drafted specs; Stage-1 work-product quality. Lowers: any resistance to re-scoping; FD slipping without a decision.

---
**Model Selection Note** — Model: claude-fable-5 · Effort: high · Status: attorney-review-ready · Stronger review: no.

> This is a Verixa patent attorney workbench output, not a final legal opinion. Qualified patent counsel and local counsel where required must verify law, facts, repo evidence, claim scope, filing authority, prosecution strategy, FTO conclusions, trade-secret risk, publication clearance, and jurisdiction-specific requirements before action.