Testing AI Category Durability With Renew

• Source: Report framework library

Encodes a Revenue AI Report framework so the agent applies the published method instead of improvising one.

1. 1.Copy the SKILL.md text below, or download the raw file.
2. 2.Create a folder named exactly testing-ai-category-durability-with-renew in your agent's skills directory.
3. 3.Save the file inside that folder as SKILL.md.
4. 4.Ask the agent one of the trigger requests below.
5. 5.Check the output against what you already know before it leaves your desk.

• Is this AI meeting-notes vendor a real category or a feature Salesforce absorbs next year
• Run RENEW on the AI SDR pitch that came in this week
• Will this vendor still exist in eighteen months, and what has to be true for that

• Should we build our own ICP model or buy a vendor for it
• Did the AI note-taker pay back its license cost over the last six months

---
name: testing-ai-category-durability-with-renew
description: Answer five questions about one AI product and produce a directional call on whether it is a durable category or a feature the platform absorbs.
---

# Testing AI Category Durability With RENEW

Answer five questions about one AI product and produce a directional call on whether it is a durable category or a feature the platform absorbs.

## When to use this skill

- An AI pitch reaches your inbox and you need a repeatable lens before booking the demo.
- A multi-year contract is on the table and the question is whether the vendor's premise survives the next foundation model release.
- An investor or board member asks whether a category is real.
- Building a shortlist and needing to separate amplification bets from replacement bets.
- Reviewing your own roadmap to check that what you are building is not a faster version of old work.

## Inputs to collect

- The product's own description of the job it does, from the vendor's site and first-call recording, in the vendor's words rather than yours.
- The named role or workflow it touches, and the headcount attached to it, from your org chart.
- How buyers currently search for or budget for this job, from search and prompt logs, procurement records, or the vendor's stated acquisition channel.
- Whether this work existed before AI, from the process owner.
- The unit economics claim, from the vendor's pricing page and any published cost-per-outcome figure.
- What the human does in the workflow after the product is in place, from the vendor's implementation plan.
- The vendor's funding, headcount, and customer references, from public filings, press, and reference calls.

## Process

Answer all five questions and record where the product sits on each axis. The test is directional rather than a rating, and it tells you what has to be true for the product to still exist in eighteen months (https://www.therevenueaireport.com/frameworks/renew).

1. R, Replace or amplify. Ask whether it makes a person obsolete or makes a person more powerful. Amplification compounds, and replacement plays lose to larger models over time. Artifact: the named role and whether the product removes or extends it.
2. E, Existing intent. Ask whether it meets buyers where they already are or has to manufacture the demand. Manufactured demand stretches cycles and inflates acquisition cost. Artifact: the evidence of existing budget or search behavior, or a note that none exists.
3. N, New work. Ask whether this work could exist without AI or is a faster version of old work. Genuinely new work creates categories, and faster old work gets compressed. Artifact: the pre-AI version of the workflow, or a statement that there is none.
4. E, Economics unlocked. Ask whether it makes something previously unaffordable affordable. Unlocked economics create new buyers, and cost reduction alone is copyable. Artifact: the buyer or use case that could not previously afford this, or a note that the claim is cost reduction.
5. W, Where does the human go. Ask where the human's remaining leverage sits in the amplified workflow. If the answer is nowhere, the seat evaporates and so does the anchor for the relationship. Artifact: the named human task that remains.
6. Write the directional call in one sentence, plus the single condition that has to hold for the product to still exist in eighteen months.
7. Record the load-bearing assumption and what would falsify it, so the call can be revisited rather than re-argued.

## Decision rules

- Weight R first. Replacement products compete directly with the next foundation model release and their advantage compresses when the model improves, while amplification products accumulate the operator's context and motion, which the model does not have (https://www.therevenueaireport.com/frameworks/renew).
- Treat a no on Existing intent as a cost signal, not a disqualifier. Products that must invent buyer intent do not scale, cycles stretch, and acquisition cost balloons (https://www.therevenueaireport.com/frameworks/renew).
- Do not accept a faster-horse answer on N. Old work with AI attached is commoditized by the next platform update (https://www.therevenueaireport.com/frameworks/renew).
- Reject cost-only bets on the second E. Cost cutting is easy to copy, unlocked economics is not (https://www.therevenueaireport.com/frameworks/renew).
- Fail the product on W if no human task remains. If the human does nothing, nothing anchors the customer relationship (https://www.therevenueaireport.com/frameworks/renew).
- Allow one rare exception to the amplification preference. A replacement product can be durable when it also creates genuinely new work rather than automating old work, and the framework states that corner is rare, carries higher risk, and has thin evidence in the Report so far (https://www.therevenueaireport.com/frameworks/renew).
- Report the answers as five positions, not a score out of five. The framework is explicit that the test is directional, not a rating, so any total invents precision the method does not have.
- Set no numeric pass bar. No page publishes one. If procurement demands a gate, use a stated rule such as no multi-year commitment where R reads replacement and W reads nowhere, and label it as a team-set rule.
- Check whether the vendor is real before the category question matters. The Gartner release behind the widely quoted forty percent cancellation forecast also estimated that only about 130 of the thousands of agentic AI vendors are real (https://www.therevenueaireport.com/research/citation-decay).
- Read a vendor repositioning as a category signal about that product, not about the whole space. The published Ledger records 11x as vendor repositioned following a TechCrunch investigation of March 24, 2025, and Artisan as vendor repositioned (https://www.therevenueaireport.com/data/reversal-ledger).
- Use the amplification target Kvarfordt states rather than a vendor's. His framing is taking work off the salesperson without replacing them, so that instead of spending 28 percent of their time actually selling they can spend 70 percent (https://www.therevenueaireport.com/frameworks/renew).
- Expect the value to sit where the work is, not where the budget is. BCG puts 70 percent of AI's potential value in core business functions including sales and marketing against 13 percent in IT, and expects agents to move from 17 percent of total AI value in 2025 to 29 percent by 2028 (https://www.therevenueaireport.com/research/what-works).
- Discount a full-autonomy roadmap. In the Capgemini series, 45 percent of organizations expect no agent involvement in the next 12 months and only 2 percent expect fully autonomous end-to-end processes (https://www.therevenueaireport.com/research/trust).

## Output requirements

- Five positions, one per letter, each with the evidence behind it.
- One directional call, stated as durable category or feature waiting to be absorbed.
- The one condition that has to hold for the product to still exist in eighteen months.
- The load-bearing assumption and what would falsify it.

Use this table shape.

| Letter | Question | Position | Evidence |
|---|---|---|---|
| R | Replace or amplify | amplify | keeps the AE, removes pre-call research |
| E | Existing intent | existing | line item already funded under sales tooling |
| N | New work | faster old work | reps did this manually before |
| E | Economics unlocked | cost reduction only | claim is hours saved per rep |
| W | Where does the human go | AE owns the call and the commitment | vendor implementation plan |

## Verification loop

1. Validate each position against the vendor's own words rather than your summary. If a position rests on your paraphrase, replace it with the vendor's sentence and re-read the position.
2. Validate R with the model test. Ask what happens to this product if the next foundation model release does the core task natively. If the product has no answer beyond being cheaper, the position is replacement rather than amplification, so correct it and continue.
3. Validate E for intent with a budget or a query. Ask which existing line item pays for this, or which query a buyer already types. If neither exists, the position is manufactured demand.
4. Validate W by naming the remaining human task in one sentence. If the sentence needs a hedge, the leverage is not real, so mark W as nowhere.
5. Re-read the five positions together and check that the call follows from them rather than from the demo impression.
6. Only proceed when all five positions cite the vendor's own words or your own process evidence, R survives the next-model test, E for intent names a funded line item or a real query, W names an unhedged human task, and the call is traceable to the five positions. Issue a recommendation only after that point. If any check fails, return the call as not yet decidable and name the missing input.

## Quality checks

- Five positions present, one per letter, none skipped.
- No score or rating is reported, because the test is directional.
- The eighteen-month condition is stated as a falsifiable statement.
- Any replacement-and-durable call carries the framework's own higher-risk and thin-evidence caveat.
- Third-party figures carry publisher and, where the source states it, sample and field date.

## Limitations

- RENEW makes a directional category call. It does not price a contract, forecast a payback period, or assess vendor solvency.
- The eighteen-month horizon is the framework's stated window, and no page publishes an accuracy rate for calls made this way.
- The replacement-and-new-work corner has thin evidence in the Report, so calls that land there carry more uncertainty than the table suggests (https://www.therevenueaireport.com/frameworks/renew).
- Vendor-supplied unit economics are claims until you see them at your own volume.
- Supporting value-distribution figures come from self-reported executive surveys with differing definitions of high performer, so read the definitions before comparing shares across sources (https://www.therevenueaireport.com/research/what-works).

## Example input

An AI product that writes and sends outbound sales emails autonomously, priced per meeting booked, pitched as a replacement for two SDR headcount. The workflow existed before AI as manual prospecting. The vendor's implementation plan has the AE reviewing a weekly summary.

## Example output

- R, Replace or amplify: replacement. The pitch names two SDR headcount removed. Replacement plays lose to larger models over time (https://www.therevenueaireport.com/frameworks/renew).
- E, Existing intent: existing. Outbound prospecting already carries budget, so demand does not need manufacturing.
- N, New work: faster old work. Manual prospecting is the pre-AI version, and faster old work gets compressed by the next platform update.
- E, Economics unlocked: cost reduction only. Per-meeting pricing is a cheaper version of an existing cost, not a previously unaffordable buyer.
- W, Where does the human go: close to nowhere. A weekly summary review is not leverage, and if the human does nothing then nothing anchors the customer relationship.

Call: feature waiting to be absorbed, not a durable category. Four of five positions read against durability, and the one that reads in favor, existing intent, is the one the framework treats as a cost signal rather than a moat.

Condition for it to still exist in eighteen months: it would have to create genuinely new work rather than automate old work, which is the framework's rare replacement-and-durable corner, explicitly flagged as higher risk with thin evidence.

Load-bearing assumption: that the outbound motion stays gateable by a human before send. Sales outreach is the worst possible fit for the condition that predicts agent success, because there is no cheap automatic check on whether an email should have been sent and sending it is irreversible (https://www.therevenueaireport.com/research/agent-reliability). What would falsify the call: a verifiable pass^4 reliability figure from the vendor plus a named buyer who could not previously afford this work at all.

## Rules of conduct

- Write for a Director, VP, or operator. Short sentences. Explain uncommon terms.
- Separate facts from assumptions. Never hide uncertainty.
- Do not invent numbers, benchmarks, quotes, or customer names.
- Do not send messages, change CRM records, or publish anything unless the user explicitly asks.
- Flag when a decision needs human review.

## Evidence

- https://www.therevenueaireport.com/frameworks/renew
- https://www.therevenueaireport.com/research/what-works
- https://www.therevenueaireport.com/research/agent-reliability
- https://www.therevenueaireport.com/research/trust
- https://www.therevenueaireport.com/research/citation-decay
- https://www.therevenueaireport.com/data/reversal-ledger
- https://www.therevenueaireport.com/frameworks/build-buy-thread

## Cite this framework

Kvarfordt, Jonathan. "RENEW." The Revenue AI Report. https://www.therevenueaireport.com/frameworks/renew

What does the Testing AI Category Durability With Renew skill do?

Answer five questions about one AI product and produce a directional call on whether it is a durable category or a feature the platform absorbs.

Where does the Testing AI Category Durability With Renew skill come from?

Report framework library. It was written by The Revenue AI Report against a 12 criterion quality rubric and graded in an independent scoring pass.

Why was the Testing AI Category Durability With Renew skill chosen for this library?

Encodes a Revenue AI Report framework so the agent applies the published method instead of improvising one.

When should the Testing AI Category Durability With Renew skill not be used?

Do not use it for: Should we build our own ICP model or buy a vendor for it Or: Did the AI note-taker pay back its license cost over the last six months

How do I install the Testing AI Category Durability With Renew SKILL.md file?

Download the file, create a folder named exactly testing-ai-category-durability-with-renew inside your agent's skills directory, and save the file inside it as SKILL.md. The agent loads it when a request matches the description.