Skills / Software-development

Adaptive Software Development (ASD)

The word 'plan' connotes false certainty. 'Speculate' honors what we actually know: that in complex systems, results emerge from the work itself, not from the plan we wrote before we started.

Adaptive Software Development (ASD) is Jim Highsmith's framework for navigating software projects where requirements and technology are fundamentally uncertain. Rooted in Complex Adaptive Systems (CAS) theory, ASD replaces deterministic planning with iterative Speculate-Collaborate-Learn cycles that treat emergent outcomes as the goal rather than a failure. Designed for teams operating in turbulent, high-change environments, ASD is one of the foundational pre-Agile methodologies and a direct precursor to the 2001 Agile Manifesto.

By Jim Highsmith  ·  Free

Use this skill See it in action

Specimen 01 · Live diagnosisAdaptive Software Development (ASD)

Input

“We're rebuilding our company's data pipeline from scratch. The stakeholders keep changing what they want out of the reports, our data engineers are…”

Diagnosis

Full transcript ↓

Calibrated referenceagent-skills.ai

The gap

Replace false certainty with Speculate-Collaborate-Learn cycles built for complex software

ASD structures every project as a repeating adaptive cycle with three phases: Speculate, Collaborate, and Learn. In Speculate, the team creates a Mission Profile — a lightweight, mission-driven document replacing the traditional project plan — and defines time-boxed feature sets for the upcoming cycle, explicitly acknowledging what is unknown. In Collaborate, concurrent component engineering and Joint Application Development (JAD) sessions drive work, with customer involvement woven into the execution rather than isolated at start and end. In Learn, formal mechanisms — quality reviews, project retrospectives, and customer focus groups — extract lessons and feed them directly into the next Speculate phase. The entire model is grounded in Complex Adaptive Systems theory: ASD treats software development as non-linear and emergent, where good outcomes result from iterative learning and adaptation rather than accurate upfront prediction.

The problem

Software projects in complex domains fail not because teams lack discipline, but because they apply mechanistic planning tools to non-linear, emergent systems. Detailed Gantt charts and fixed requirements documents create the illusion of control while hiding the true nature of complex software: that requirements surface through collaboration, that technology constraints reveal themselves during implementation, and that the most valuable insights arrive after the work begins. ASD was created specifically to solve this mismatch — treating software development as a Complex Adaptive System where the goal is not prediction, but adaptive response.

The solution

Stop building the wrong thing confidently. Use ASD's Speculate-Collaborate-Learn cycles to replace false certainty with structured adaptation — so your team can navigate genuine complexity, extract real learning from each cycle, and deliver outcomes that reflect what you actually discover, not what you guessed at project kickoff.

You bring

A software project context — its domain, team, key uncertainties, and stakeholders
Your current planning approach and where it is breaking down
A sense of which requirements or technical decisions feel most unstable
Willingness to reframe 'we don't know yet' as a legitimate project state, not a failure

You get

A Mission Profile document — a mission-driven, uncertainty-aware project foundation that replaces a rigid project plan
A structured Adaptive Cycle Plan with time-boxed feature sets and explicit risk acknowledgment
A facilitated Speculate-Collaborate-Learn retrospective format your team can run at each cycle end
A CAS-grounded diagnostic for why your project feels chaotic — and how to work with emergence rather than against it

In action

Watch the methodology work.

Three specimens from a single real session: the same situation, before and after, the full transcript, and the skill answering live in the channel where the work happens.

Fig.01  —  The same situation, two ways

Specimens A · B

Specimen ABefore

Without the skill

You have a 12-data-source pipeline rebuild with a Q3 deadline. You've written three project plans in five weeks. Each one became obsolete within days as engineers discovered new data quality issues and the finance team changed their report requirements after seeing the first prototype. Your standups are dominated by replanning rather than progress. Stakeholders are asking why the plan keeps changing.

×The situation as it arrives — before the methodology is applied.

Specimen BAfter

With Adaptive Software Development (ASD)

Your project runs on four-week Adaptive Cycles, each anchored by a Mission Profile that everyone understands. At the start of each cycle, you speculate on a specific feature set, explicitly document what you don't know, and define what you'll learn by the end. At cycle end, a structured customer focus group with the finance team generates real requirements from working software — and those insights become the input to your next Speculate phase. Stakeholders stop asking why the plan changed because they understand you're running adaptive cycles, not executing a prediction.

◦The same situation once the methodology holds.

Fig.02  —  A real session, unedited

In-app · assistant

Adaptive Software Development (ASD)Claude

Adaptive Software Development (ASD)

I work from Jim Highsmith's Adaptive Software Development framework — the Speculate-Collaborate-Learn cycle built for projects where the plan you write on day one will definitely be wrong by week three. Before we dive in, tell me: what's the nature of the project you're working on, and where is the uncertainty hitting you hardest? Is it unclear requirements, shifting technology, team coordination, or something else?

We're rebuilding our company's data pipeline from scratch. The stakeholders keep changing what they want out of the reports, our data engineers are discovering new constraints every sprint, and I feel like I'm replanning constantly. Every two weeks I write a new project plan and within days it's already wrong.

That's a textbook Complex Adaptive System — your requirements are emergent (stakeholders discover what they want by seeing working output), and your technical constraints are self-revealing (the engineers learn what's actually possible by doing the work). Writing a detailed plan in that environment isn't discipline — it's wishful thinking. The first thing ASD would have you do is stop calling what you write a 'plan' and start calling it a 'Mission Profile.' Not a semantic trick — a real shift in what you're committing to. A Mission Profile says: here is our mission, here are the feature sets we're speculating on for this cycle, and here is what we don't yet know. Can you tell me what you actually know is true about this project right now, versus what you're guessing?

What I know: we need to ingest from 12 data sources, the finance team is our primary customer, and we have a hard deadline of Q3 for the first usable version. What I'm guessing: what the finance team actually needs to see in the reports, and which data sources will be clean enough to use without significant transformation work.

Perfect — that's exactly the foundation for your Mission Profile. The mission is clear: a data pipeline serving finance by Q3. Your 'what we know' becomes the fixed elements of your Adaptive Cycle structure. Your two unknowns — what finance actually needs, and which data sources are tractable — become your highest-priority learning goals for your first Speculate cycle. In ASD, you'd deliberately front-load the riskiest and most uncertain feature sets into early cycles, not because you'll solve them, but because you want to fail fast on your most dangerous assumptions. Here's what I'd propose: instead of planning all 12 data sources, your first Adaptive Cycle plan speculates on 2-3 sources and one working report for finance — with the explicit goal of running a customer focus group at the end to find out what the finance team actually learns from seeing real output. That feedback becomes the input to your next Speculate phase. Want me to draft a Mission Profile for this project based on what you've shared?

Yes — but I'm worried about how to present this to my stakeholders. They expect a traditional project plan with all the milestones laid out.

That's where the Mission Profile earns its keep. It's not a blank page — it has structure. It includes: the project mission statement, the customer profile (finance team, their key outcomes), the adaptive cycle structure (how many cycles, rough time-boxes), a feature-set list showing what you're speculating on and in what order, a risk and uncertainty log (what you know you don't know), and success criteria for the first cycle. Stakeholders get a real document — they just get one that's honest about uncertainty rather than falsely confident. Many executives actually prefer it once they see it, because it stops the cycle of 'why is the plan wrong again?' and replaces it with 'here is what we learned and here is what we're doing next.' Let me draft the Mission Profile structure for your data pipeline project. I'll build it from the facts you've given me.

This is starting to make sense. The reason my plans keep failing is that I've been writing plans for a deterministic project when I actually have a complex adaptive one.

Exactly — and that's Highsmith's core insight from the original ASD book. You weren't doing planning wrong; you were using the wrong tool for the system you're actually dealing with. The shift from 'plan' to 'speculate' isn't lowering your standards — it's being honest that in a Complex Adaptive System, the best information you'll ever have about what to build next comes from the last cycle of work, not from an upfront requirements document. Here is a draft Mission Profile structure for your data pipeline rebuild...

Reply to Adaptive Software Development (ASD)…

What the skill is doing

01Mission Profile Generator

02Adaptive Cycle Planner

03Speculate-Collaborate-Learn Retrospective Facilitator

Fig.03  —  The same skill, in the channel you already use

WhatsApp · live

9:41

ASAdaptive Software Development (ASD)typing…

Messages are end-to-end encrypted. No one outside of this chat can read them.

Today

We're rebuilding our company's data pipeline from scratch. The stakeholders keep changing what they want out of the reports, our data engineers are di…9:14✓✓

Before we dive in, tell me: what's the nature of the project you're working on, and where is the uncertainty hitting you hardest?9:14

Is it unclear requirements, shifting technology, team coordination, or something else?9:15

Message

The same skill, where the work happens.

No new app to learn. The methodology runs over the WhatsApp Business API, so the answer lands as a reply in the thread you’re already in — same rigour, zero context-switch.

◦Reads the situation, names the pattern, returns one concrete next move.

◦Delivered in seconds, inside a conversation that already exists.

Specimen · WhatsApp Business API · live

Capabilities

What it does, specifically.

Each capability is a distinct move drawn straight from the source methodology — not a generic assistant guessing.

CapabilityC-01

Mission Profile Generator

Guides you through creating a Mission Profile — ASD's replacement for the traditional project plan. Unlike a project plan, a Mission Profile defines the project's purpose, target customers, key success criteria, time-boxed cycle structure, and known uncertainties without pretending to know the unknowable. It becomes the anchor document for all Speculate phases.

◦ MethodBased on Highsmith's Mission Profile artifact from ASD (2000), which replaces the project charter and Gantt chart with a mission-driven, uncertainty-honest project foundation.

CapabilityC-02

Adaptive Cycle Planner

Helps you design a time-boxed Adaptive Cycle Plan: select feature sets for the upcoming cycle, sequence them by risk and dependency, and define clear learning goals for each cycle — not just delivery targets. Each cycle is treated as a controlled experiment, not a commitment to a predetermined outcome.

◦ MethodBased on ASD's Speculate phase concept of feature-set-based, time-boxed iteration — where the word 'plan' is explicitly replaced with 'speculate' to honor project uncertainty.

CapabilityC-03

Speculate-Collaborate-Learn Retrospective Facilitator

Structures your cycle-end reviews using ASD's formal learning loops: quality reviews (what did we build and how well), project retrospectives (how did we work together), and customer focus groups (what did the customer actually learn from the working software). Outputs feed directly into the next Speculate phase.

◦ MethodBased on Highsmith's 'Learn' phase mechanisms — quality reviews, retrospectives, and customer focus groups — designed as structured knowledge-extraction events rather than informal debriefs.

CapabilityC-04

CAS Complexity Diagnostic

Uses Complex Adaptive Systems theory to diagnose why your project is struggling: Are you applying linear process tools to a non-linear problem? Is command-and-control leadership suppressing the collaborative emergence your team needs? The diagnostic maps your project's symptoms to specific ASD principles and suggests which phase of the SCL cycle needs attention.

◦ MethodBased on Highsmith's application of CAS theory to software projects — the core theoretical foundation of ASD distinguishing emergent from deterministic systems.

CapabilityC-05

Collaboration vs. Command-Control Coach

Helps team leads and project managers shift from directive, hierarchical control toward the collaborative, self-organizing patterns ASD calls for. Covers Joint Application Development (JAD) session design, concurrent component engineering coordination, and how to involve customers as active collaborators rather than external approvers.

◦ MethodBased on ASD's Collaborate phase emphasis on concurrent engineering, JAD sessions, and Highsmith's principle of 'collaboration over command-and-control' as a prerequisite for adaptive systems.

Tested

Graded before it shipped.

Every skill is scored against independent scenarios for methodology fidelity before it goes live — not vibes, a rubric.

What it produces

OutputD-01

Mission Profile

A lightweight project foundation document that defines mission, target customers, success criteria, known risks, and cycle structure — without pretending to know what will only be learned through the work itself. Replaces the traditional project plan or project charter.

OutputD-02

Adaptive Cycle Plan

A time-boxed feature-set breakdown for the current cycle — not a Gantt chart, but a speculative map of what the team will attempt, what risks they are accepting, and what learning goals they are pursuing. Revised at the start of every cycle.

OutputD-03

SCL Retrospective Report

A structured cycle-end learning document capturing quality review findings, team retrospective insights, and customer focus group feedback — formatted to flow directly into the next Speculate phase as actionable input.

OutputD-04

CAS Project Diagnostic

A written diagnostic that maps your project's challenges to Complex Adaptive Systems principles — identifying where deterministic assumptions are creating friction and where adaptive practices could replace them.

The source

Grounded in the original work.

Every answer traces back to a real source and the practitioner who wrote it — not a secondhand summary. Here is the source of record.

Source authorA-01

Jim Highsmith

Jim Highsmith is a founding architect of the Agile movement — a signatory of the 2001 Agile Manifesto and the originator of Adaptive Software Development (Dorset House, 2000). He also authored Agile Project Management (Addison-Wesley, 2004; 2nd ed. 2009) and served as a Fellow at the Cutter Consortium and director at ThoughtWorks. Highsmith spent decades coaching executive teams and software organizations through high-uncertainty projects, and his CAS-grounded critique of deterministic planning directly shaped the principles behind the Agile Manifesto.

Status · Inspired by Jim Highsmith’s work — not yet claimed. Are you Jim Highsmith?

Primary sourceS-01

Adaptive Software Development: A Collaborative Approach to Managing Complex Systems (2000)

by Jim Highsmith

Agile Manifesto co-signatory; author of Adaptive Software Development (2000) and Agile Project Management (2004/2009); former Fellow, Cutter Consortium; director, ThoughtWorks

Read the original ↗

Citationjimhighsmith.com

In the build queue

Be first to run it.

Adaptive Software Development (ASD) is being built right now. Leave your email and we’ll tell you the moment it goes live.

Notify meEmail

At launchWe're six weeks into a complex software rebuild and I've already written four project plans — each one wrong within days. I think we need to stop planning and start speculating. Can we build a Mission Profile and design our first Adaptive Cycle?