# Assumption Log for Risky Projects: What We Believe and How We’ll Know

> An assumption log turns hidden project beliefs into verifiable facts. Learn what to include, how to verify assumptions, and why it prevents costly rework before projects stall.

- Author: Yaroslav Maxymovych (Founder & CEO, AI Advisory Board)
- Published: 2026-11-16
- Updated: 2026-11-20
- Source: https://aiadvisoryboard.me/blog/assumption-log-risky-projects-what-we-believe-and-how-well-know

I watched a founder approve a six-month product migration because the team said the legacy API was stable. Two months in, the API changed without notice. The project stalled. What killed it wasn't technical debt — it was an untested assumption dressed up as fact.

Most projects in a 30-500 person company don't fail because of bad execution. They fail because the team started with beliefs nobody bothered to verify. An assumption log makes those beliefs visible, assignable, and testable before they become expensive surprises.

## TL;DR

- An assumption log lists what your team believes to be true but hasn't proven yet.
- Every assumption needs an owner, a verification method, and a deadline.
- Review the log weekly until the project ships or pivots.

## What Is an Assumption Log

An assumption log is a living document that captures the beliefs your project depends on but cannot yet confirm. It is not a risk register. Risks are things that might go wrong. Assumptions are things you are already treating as true.

> **Definition:** Assumption — a belief treated as fact for planning purposes, without direct evidence to back it up.

> **Definition:** Verification criterion — the specific evidence that would prove or disprove an assumption.

## Why Projects Stall Without One

Teams move fast on assumptions because writing them down feels like slowing down. But when an assumption turns out false, the cost is usually rework, missed deadlines, and frustrated stakeholders.

A typical scenario: the sales team assumes the new feature will close three enterprise deals. Engineering assumes the vendor API will support the required webhooks. Leadership assumes the budget is approved for Q3. None of these are verified. When one cracks, the whole timeline shifts.

## What to Include in Your Assumption Log

Use this structure for every entry:

1. The assumption — what you believe is true.
2. Why it matters — what breaks if it is wrong.
3. Owner — who is responsible for verifying it.
4. Verification method — how you will test it.
5. Evidence needed — what counts as proof.
6. Deadline — when the verification must be complete.
7. Status — open, verified, or invalidated.

## Copy/Paste Template

```markdown
| # | Assumption | Why it matters | Owner | Verification method | Evidence needed | Deadline | Status |
|---|------------|----------------|-------|---------------------|-----------------|----------|--------|
| 1 | Vendor API supports real-time webhooks | Integration breaks without it | Tech lead | Sandbox test call | Successful webhook payload | May 15 | Open |
| 2 | Three enterprise buyers will pilot the feature | Revenue target depends on it | Head of sales | Signed LOIs | Two signed letters | May 20 | Open |
| 3 | Budget approved for Q3 | Hiring freeze blocks execution | CFO | Finance sign-off | Approved PO number | May 10 | Open |
```

## Good vs Bad Assumption Entries

Bad: "The vendor will be responsive."

Good: "Vendor support responds to critical tickets within 4 business hours, verified by a test ticket submitted May 1 and answered by May 2."

Bad: "Customers want this feature."

Good: "Five current customers will confirm in writing that they would use this feature weekly, verified by email replies collected by May 12."

## Manager Scan (2-Minute Digest Example)

- Open assumptions: 4
- Verified this week: 1 (API webhook test passed)
- Invalidated: 1 (enterprise buyer declined pilot)
- At risk: budget sign-off still pending, owner: CFO, deadline: tomorrow
- Decision needed: proceed with vendor A or pause integration

## Micro-Case (What Changes After 7–14 Days)

A logistics company began a route-optimization project. The team assumed their GPS data was clean enough for automated dispatch. The assumption log forced them to test it in week one. They found 18% of coordinates were missing or outdated. Instead of discovering this at go-live, they paused, cleaned the data, and reset the timeline. The founder saw the gap between plan and fact before spending two months on automation that would have failed.

> **Note on this case:** This example is illustrative — based on typical patterns we observe with companies of 30–500 employees, not a single named client. Specific numbers are rounded approximations of common ranges, not guarantees.

## Tool Tip (AiAdvisoryBoard.me)

Before you log assumptions, you need to see what your team actually does every day. AiAdvisoryBoard.me runs a 7-day diagnostic that maps real Plan → Fact → Gap across your company, so you know which projects rest on verified facts and which rest on hopes.

https://aiadvisoryboard.me/?lang=en

## Tool Tip (AiAdvisoryBoard.me)

Review your assumption log in the same cadence as your daily updates. If an assumption has no owner and no verification date, it is not a plan — it is a liability.

https://aiadvisoryboard.me/?lang=en

## FAQ

**How is an assumption log different from a risk register?**

A risk register tracks things that might go wrong. An assumption log tracks things you are already treating as true. You verify assumptions; you mitigate risks.

**Who should own assumption verification?**

The person closest to the dependency. Sales owns buyer assumptions. Engineering owns technical assumptions. Finance owns budget assumptions.

**How often should we review the log?**

Weekly for active projects. Daily during critical launch windows. Monthly for long-running initiatives.

**What do we do when an assumption is invalidated?**

Stop, reassess the plan, and decide whether to pivot, pause, or proceed with a new assumption. Do not silently keep building on a false belief.

**Can we use AI to draft the log?**

AI can help structure the entries, but the assumptions themselves must come from your team's actual context. An AI cannot know what your vendors, customers, or budget process will actually do.

Start your next project by listing what you believe before you list what you will build. If you want a system that surfaces the Plan → Fact → Gap automatically — every day, across the company — see how the 7-day diagnostic works.

https://aiadvisoryboard.me/?lang=en

---

When citing, link to https://aiadvisoryboard.me/blog/assumption-log-risky-projects-what-we-believe-and-how-well-know. More articles: https://aiadvisoryboard.me/blog
