Why do hourly bids lose now?
Software now ships three ways: hand-written, AI-assisted and agentic, often on the same project. The client does not pay for who or what wrote the code. It pays for what the work cost and how well it held up.
- An agent has no hours to bill. On a time-and-materials contract, every hour that AI saves is an hour that you do not bill.
- On a fixed price, AI cost is your risk. A frontier model on a trivial ticket, or seats that nobody uses, comes out of your margin.
- The client cannot compare bids. Two firms quote different hours for the same scope, and neither can show what an hour delivers.
How do you win the bid on price?
- Quote in points. Scope the work with the client, then quote in delivery points and a price per point.
- Price below the hourly bid. Your AI savings let you offer a lower price per point than a firm that bills hours, and still hit your target margin.
- Show the evidence. Pensero scores each pull request and ticket as it ships, and the client sees the points behind each invoice.
- Bill the accepted points. The client pays for delivered work, not for time.
| Hourly bid | Delivery point bid | |
|---|---|---|
| What the client buys | Time | Delivered work |
| What happens to AI savings | Lost revenue, or hidden | Shared: lower price for the client, margin for you |
| Can the client compare bids | No, hours are not equal | Yes, the same unit for every bidder |
| Proof for the client | Timesheets | Points traced to each pull request and ticket |
How do you keep the margin?
Read AI spend per point by client, team and model. When an agent delivers a point for less, the difference is your margin. When a model costs too much for the type of work, you see it before it reaches the P&L.
How do you prove the transformation in the proposal?
The client asks how much of the work really moved to AI agents. Show it week by week, by type of work, instead of a slide that says "AI-enabled". Add defects and rework, on agent and human code alike, so the client sees that quality holds as volume goes up.
How do you win the renewal?
Use the delivery point history of the engagement as the baseline for the next bid: points delivered, price per point, quality and rework. You re-compete on a track record, not on a discount.
Common mistakes
- Billing agent hours. Runtime is not delivered work.
- Using story points as the price unit. The team that bills them also sets them, so the client cannot compare vendors.
- Hiding AI use. The client finds out, and the savings become a discount instead of a shared gain.
- Measuring output without quality. More points with more defects lose the client's trust.
How Pensero helps outsourcing firms bid
- Pensero scores every pull request and ticket in delivery points, the same way for engineers, AI-assisted engineers and AI agents, and traces each point to the work behind it.
- Pensero does not deliver the work and does not grade one vendor for another. It only measures.
- Your data stays private. You decide what each client sees.
- Read how to price software outsourcing when AI agents write the code, the KPIs for an outsourcing contract and how IT services firms protect their margin.
- Read what a delivery point is, whether to bill clients for AI agent hours and how to price a software project in 2026.
- Book a scoping session to price your next bid in delivery points.
Sources
- The Wall Street Journal, 13 November 2026: Baker McKenzie Looks to Abandon the Billable Hour as Default.
- Bloomberg Law, 16 November 2026: Wake Up Call: Baker McKenzie Nods to AI in Hourly Billing Shift.
- Nikkei Asia, 17 November 2026: AI forces Vietnam's FPT Software to shift from hourly billing model.
Frequently asked questions
How can an IT outsourcing company win RFPs now that AI writes the code?
Quote per delivery point instead of per hour. Pass part of the AI savings to the client as a lower price per point, keep the rest as margin, and show the client the evidence behind each point.
How do I quote a software project without man-hours?
Scope the work, quote in delivery points and a price per point, and bill the accepted points after the work ships. A delivery point is magnitude × complexity, scored after the work ships.
How do I keep margin on a fixed-price contract with AI?
Read AI spend per point by client, team and model, and send trivial work to cheaper models. The margin is the difference between your price per point and your cost per point.
Who sets the price per delivery point?
The firm. The delivery point measures delivered work, not price, so each bidder quotes its own price and competes on price, quality and reliability.
