---
title: Absorb → Innovate → Disrupt
synced_from_vault: true
vault_source: 03-living-docs/patterns/Absorb-Innovate-Disrupt.md
public: true
name: Absorb → Innovate → Disrupt
description: >-
  Every new technology gets deployed in three stages. (1) Absorb — do the old
  work better. (2) Innovate — do things only possible with the new tech. (3)
  Disrupt — redefine the question entirely. Most organizations stop at stage 1
  and mistake it for "doing AI."
type: pattern
category: strategy
origin: >-
  Benedict Evans, "AI eats the world" (May 2026 deck). Echoes the historical
  pattern of every prior platform shift: mainframes, PCs, web, smartphones —
  same three stages each time.
created: 2026-05-18T00:00:00.000Z
tags:
  - pattern
  - ai
  - strategy
  - technology-adoption
  - platform-shifts
---

> **Core pattern:** Every new general-purpose technology gets deployed in three stages. Skipping ahead is impossible; staying in stage 1 forever is the most common failure mode.

---

## The three stages

| Stage | What it looks like | Example (spreadsheets) | Example (AI today) |
|------|--------------------|------------------------|--------------------|
| **Absorb** | Do the old work, with the new tool. The deliverables look the same; production gets cheaper or faster. | Accountants stop using ledgers and 10-keys; output is still the financial statement. | "Use AI to write our existing campaigns faster." Generated email drafts. Sales-call summaries. |
| **Innovate** | Do things that were previously impossible because the cost was prohibitive. New deliverables emerge from new questions. | Financial modeling, scenario planning, what-if analysis — none of which existed as workflows before the spreadsheet. | "Analyze every customer call we've ever had to find the churn signals" — see [What-Was-Impossible-Now-Cheap](/patterns/what-was-impossible-now-cheap). |
| **Disrupt** | Redefine the question. The category itself changes. New companies arrive whose business doesn't make sense in the old framing. | Decision support, FP&A as a profession, the entire venture-backed SaaS thesis (subscription P&Ls). | TBD. Likely vertical-AI businesses that sell *outcomes* not tools. Possibly browser/voice/agent UX shifts. |

---

## Why most orgs get stuck at Absorb

Absorb-stage AI is **symmetric**: every competitor does it. Faster campaigns, summarized calls, generated drafts. It's table stakes — necessary but not strategy. It feels productive ("we shipped 14 AI features!") and looks good in board decks, but produces no defensible advantage because everyone ships the same thing.

The stickiness of Absorb has three causes:
1. **The org chart only has roles for the old work.** Innovate-stage work needs roles nobody has hired for.
2. **Customers can describe Absorb work** ("make my existing campaigns faster") and can't describe Innovate work ("what previously-impossible question would change your business?"). Roadmap-by-customer-request stays in Absorb forever.
3. **It's measurable.** Absorb has obvious KPIs (time saved, cost reduced). Innovate has unclear KPIs because the deliverable doesn't exist yet.

---

## How to tell which stage you're in

For any AI capability you've shipped or are planning, ask:

- **Absorb test:** Could a competitor ship the same thing in three months? If yes, it's Absorb. Useful, but not strategy.
- **Innovate test:** Does this answer a question that didn't appear in your customer's last RFP? If yes, you're entering Innovate.
- **Disrupt test:** Does this make a category of existing vendor obsolete? If yes, you might be the disruptor — or you might be wrong. Either way, you're in territory that matters.

---

## The "spreadsheet for an accountant vs a lawyer" test

Evans's framing: imagine showing the first spreadsheet to two professionals.

- **Accountant:** instant clarity. "This is my job." Spreadsheets absorb the obvious task and then Innovate beyond it for the next 40 years.
- **Lawyer:** "Very cool, but..." The tech is real, the use case is unclear, the *job* doesn't decompose into rows × columns. Spreadsheets stay irrelevant to law for a generation.

For AI in 2026, the question is: *am I the accountant or the lawyer?* Some industries see obvious Absorb wins (software engineering, customer support, marketing copy). Some industries are still squinting at the lawyer-side of the line. The clarity of stage-1 fit is itself a signal.

---

## Failure modes

- **Declaring victory at Absorb.** "We're an AI company now" because 30% of your engineers use Copilot. This is hygiene, not strategy.
- **Skipping to Disrupt without Absorbing.** Building the speculative future product while the existing product rots. Internal credibility collapses; external customers don't follow.
- **Stage-mixing in one team.** Innovate work doesn't survive on an Absorb team's roadmap — it gets cut every sprint because it lacks an Absorb-shaped KPI.

---

## Cross-References

- Companion frames: [Was-Cost-of-Task-Your-Moat](/patterns/was-cost-of-task-your-moat), [Task-vs-Job](/patterns/task-vs-job), [What-Was-Impossible-Now-Cheap](/patterns/what-was-impossible-now-cheap)
- Related patterns: [AI-Planning-Inflection-Point](/patterns/ai-planning-inflection-point), [Augmentation-Over-Automation](/patterns/augmentation-over-automation), Information-Diet-Asymmetry