You are deciding how to structure a pilot so it ends in a purchase order. Write the paid contract before the pilot starts. Charge for the pilot, sign two or three success metrics with the economic buyer, put the price and the conversion date in the pilot paper, and start security and procurement in week one. At readout you check the boxes and ask for the signature. SaaStr readers report paid pilots convert far more often than free ones: 74% vs 42%.
Pilot, paid POC, or straight contract
| Motion | Use it when | Watch for |
|---|---|---|
| Straight annual contract with an opt-out | You have a brand and references. Lemkin: established vendors flip pilots into contracts the buyer can cancel after 60 days | Early-stage buyers may still insist on proof first |
| Paid pilot that converts to a priced contract | A new vendor selling into a buyer who needs proof inside their own workflow. SaaStr: enterprise buyers often won’t buy from a startup without one | Selling it twice: once to start, again to convert (Lemkin) |
| Paid proof of concept on one technical question | The open risk is integration or accuracy, and a demo on their data cannot settle it | A POC that proves the tech and stops short of the budget owner |
| Demo on the buyer’s own data | AI products where value shows in the first meeting. a16z: demos are taking over the classic POC role | Treating a good demo as a closed deal |
| Free pilot | Self-serve products that deploy in minutes (Lemkin separates freemium from free pilots) | Real deployment work and process change. Lemkin: free pilots of that kind almost never convert |
AI raised the bar. Gartner forecast that companies would abandon at least 30% of generative AI projects after proof of concept by the end of 2025, citing data quality, risk controls, cost, and unclear business value. a16z adds that many enterprise buyers want to test an AI category once. If your pilot fails, you may not get a second try at that account.
Write the conversion before kickoff
| Clause | What to write | Why |
|---|---|---|
| Problem and scope | One team, one workflow, named in-scope and out-of-scope items | SaaStr: be specific to avoid scope creep |
| Success metrics | 2–3 numbers with a baseline, an owner, and how you measure them | SaaStr; Lemkin: without them the pilot becomes an endless evaluation |
| Clock | Start date, a 30-day review, an end date | SaaStr’s worked example: 60 days, review at 30 |
| Price and conversion | Pilot fee, the annual price if metrics hit, and the signing date | Session judgment: the PO is the goal; agree the price you need before you prove anything |
| People | Day-to-day lead, decision-maker, and who signs the PO | SaaStr: name the decision-maker on their side and a dedicated owner on yours |
| Path to production | Security review, data access, procurement steps, and who owns each | a16z: buyers expect compliance in place from the start |
Not legal advice. Have counsel review the pilot terms and the order form.
Work backward from the contract you need. Ask the buyer before kickoff: if we hit these numbers, will you buy, at what price, and who signs? A pilot that succeeds against the wrong price is a lost deal. That is judgment from a FounderNexus session, not a survey. Price the pilot so the buyer can approve it fast, and save the larger number for the contract.
Map the buying group in week one
Gartner’s 2023 B2B Buying Report counts 5–11 stakeholders in the average enterprise buying group, drawn from about 5 business functions. Write each one down with a stance: for, against, or neutral.
| Role | What they need from the pilot | Start them |
|---|---|---|
| Champion | Proof they can show their boss | Before kickoff |
| Economic buyer | A business case in cost, risk, or revenue terms | Before kickoff. The success metrics are theirs |
| Users | A quick win in their own workflow | Week one. Skeptics turn when they shape the pilot (session judgment) |
| Security and IT | Answers on data use, logging, and controls | Week one. a16z: AI buyers ask whether you train on their data and how you log prompts and outputs |
| Procurement | Vendor forms, paper, and the PO path | Week one, in parallel. Finding the path at readout opens a second negotiation |
A pilot with one contact dies when that person changes jobs. Keep at least two threads alive above the champion.
Run the pilot like a sale
Pilot to signed contract
- Scope one problem. One team and one workflow. Write what is out of scope so new requests wait for the contract (SaaStr).
- Sign the success metrics. 2–3 numbers the economic buyer owns, with baselines (SaaStr). Instrument them in the product where you can.
- Paper the conversion. Pilot fee, contract price, term, and the signing date go in the pilot agreement. Start security and procurement the same week.
- Run it with a mid-point review. Show a visible win early. Hold a formal check against the metrics at about day 30 of a 60-day pilot (SaaStr example).
- Read out against the numbers. Show each metric against its baseline. Turn the result into an ROI case finance can sign.
- Ask for the signature. These were the criteria; you hit them; here is the order form. If you did steps 1 to 3, this meeting is short.
a16z notes 70% of buyers in its enterprise survey rank speed of deployment as a top factor. A long pilot works against a buyer who wants results this quarter.
When the pilot stalls
| Stall | Move |
|---|---|
| No economic buyer in the room | Go get the person who owns the budget. An unfunded pilot does not convert |
| Metrics too vague to close against | Rewrite them as numbers with baselines and get them re-signed |
| Scope keeps growing | Move new asks to the contract scope. Out-of-scope items wait for the signature |
| Champion goes quiet | Use the second thread. Ask the economic buyer for the readout date |
| Procurement takes over the timeline | Ask the champion or buyer to push, tied to the date that forced the pilot in the first place |
| They like it but have no budget | Shrink the first contract to one team or site before you discount. Session judgment: a discount sets the price for the next deal |
Judgment from a FounderNexus session: cut scope before you cut price, and ask for the PO the day you hit the metrics. A pilot that ends with “the team liked it” has failed as a sale.
Pair with pricing and sales hiring
Pick seats, usage, or outcomes on seat vs usage vs outcome pricing for AI SaaS. Lead the pilot pitch with the job and the outcome on how to market an AI product without leading with AI. If pilots pile up without an owner for the close, read when to hire the first VP of Sales. Orient the rest of the motion from the GTM hub.
fn-content has no verified pilot-conversion atom yet. Tracked as benchmark request: pilot-to-paid conversion rate. Until then, cite the named SaaStr, a16z, and Gartner figures above. Do not quote a blended market rate.
Sources
- SaaStr, The Top 20+ SaaStr Surveys On Freemium, Discounts, Paid Pilots, Raising Prices, And Much More — Jason Lemkin, 27 May 2022. Reader polls: 42% say free pilots generally convert to paid; 74% say paid pilots convert at the end of the pilot. Sample size not stated.
- SaaStr, Dear SaaStr: What is The Typical Conversion from Paid Pilot to Annual Contract in SaaS? — Jason Lemkin, 25 October 2021. 70%+ once you do it right; about 60% to over 90% across companies he works with; established vendors flip pilots into contracts with a 60-day opt-out; more process change means lower conversion.
- SaaStr, Dear SaaStr: Should We Charge for Pilots and How Much? — Jason Lemkin, 31 January 2023, updated 1 September 2026. Charge for any true pilot; a free pilot is an extended demo; free pilots with real deployment work almost never convert; freemium is a different case.
- SaaStr, Dear SaaStr: We’re Starting a Big Paid Pilot. How Do Maximize the Chances of Success? — 29 May 2025. Define 2–3 key metrics; specific scope to avoid creep; 60-day example with a 30-day review; name the decision-maker; tie the pilot to expansion.
- SaaStr, Dear SaaStr: Why Do Enterprise Customers Insist on Pilots? — Jason Lemkin, 5 May 2025, updated 4 September 2026. Pilots get sold twice; many enterprise buyers will not buy from a startup without one; set success criteria up front.
- a16z, Need for Speed in AI Sales — Seema Amble and James da Costa, 21 November 2025. a16z Enterprise survey: 70% of buyers rank speed of deployment a top factor; 57% expect positive ROI within three months; 11% immediately. Demos taking over the POC role; compliance expected from the start; buyers want to test an AI category once.
- Gartner, Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept By End of 2025 — 29 July 2024. At least 30% abandoned after POC, from poor data quality, inadequate risk controls, escalating costs, or unclear business value. A forecast, not a measured rate.
- Gartner, B2B Buying Report — 2023. Average enterprise buying group of 5–11 stakeholders from an average of 5 business functions.
- FounderNexus session — Operator judgment: the purchase order is the goal; agree the price before you prove anything; map approvers, payers, and blockers; start security and procurement in parallel; cut scope before price; ask for the PO at readout. Not a survey.
Founders who have turned enterprise pilots into purchase orders will pressure-test your success criteria, buyer map, and conversion clause in a FounderNexus session.