SA Module 05: Your AI-Augmented SA Practice

1

What you've built across this pathway

The through-line of this pathway is stated in Module 01: AI generates pattern answers; the SA owns every constraint, tradeoff, and decision the pattern doesn't account for. Everything else — the options analysis discipline, the ADR constraint documentation, the integration architecture specifics, the communication and governance practices — serves the SA's role as the person in the engagement who is accountable for the architecture being right, not just looking right.

🗺️

Module 01 — Architecture at Speed

AI generates pattern-level options fast. The SA adds the constraint layer — system limits, regulatory requirements, operational capability, SLA commitments — that makes generic patterns specific. Architecture ownership means being accountable for every design decision, including its gaps.

📋

Module 02 — Design Decisions

ADRs carry architectural knowledge forward past the SA's engagement. Complete consequences — enables, constrains, requires — prevent the specific class of production failure that comes from someone changing a design without knowing what it depended on. AI drafts; SA adds what only the engagement context contains.

🔌

Module 03 — Integration Architecture

Guidewire integration patterns, data migration architecture, legacy system connectivity — all require constraint discovery that AI can't do. Pattern selection follows constraint application; data quality realities shape migration approach; open items must be explicit in architecture presentations.

🎤

Module 04 — Communication & Governance

Different audiences need different architecture translations — AI produces these efficiently; SA verifies accuracy and consistency. Proactive gap acknowledgment in ARBs is stronger than reactive defence. When overruled: document the recommendation, then execute the decision fully.

2

What an AI-augmented SA day actually looks like

Starting an integration design

Options generation in 20 minutes

Describe the integration requirement and constraints to AI; get 3 viable pattern options with tradeoffs. Spend the time on applying the constraint layer — system capabilities, regulatory requirements, team skills — rather than generating the option space from scratch.

After each significant decision

ADR in 15 minutes

Bullet your decision notes — what, why, alternatives, constraints. AI structures the ADR. You add what only you know: verbal commitments with confidence levels, time-bounded constraints, the real reasons alongside the official ones. Consequences section: enables, constrains, requires — not just benefits.

Before any architecture presentation

Challenge your own design

Run an explicit review prompt: single points of failure, NFR gaps, constraint violations, what a critical reviewer will raise. Find the weak points before the audience does. Proactive acknowledgment of known gaps is stronger than reactive defence in an ARB or CTO presentation.

When communicating to stakeholders

Audience-adapted versions

AI adapts the same architecture content for CTO, development team, PM, and steering committee. You verify every version for accuracy — especially that strategic risks appear where decisions are being made, and that open items are explicit rather than footnoted.

During data migration planning

Framework, then real constraints

AI generates the migration framework quickly. You apply the real data quality findings, the actual legacy system state, the fixed regulatory deadline. Data quality issues are architecture decisions, not just risks. Let the data reality shape the approach — don't fit the data into a predetermined plan.

When challenged or overruled

Reasoning-based response

Engage challenges by understanding them fully before responding. Hold positions with constraint evidence, not seniority. When overruled: make the case once more with risk quantified, document it, then execute the decided approach with full commitment. Professional authority is maintained through reasoning, not position.

3

Positioning your AI capability — the senior SA rate conversation

Generic positioning

"I have Guidewire solution architecture experience and I use AI tools to help with architecture analysis and documentation."

Premium positioning

"I use AI systematically across the SA lifecycle — options generation for integration patterns, Architecture Decision Record drafting, constraint analysis across regulatory and system dimensions, and audience-adapted communication for CTOs through to development teams. The AI generates the pattern space and the structure; I apply the constraint layer that makes generic patterns specific to each client's environment. For Guidewire implementations specifically, I'm rigorous about the constraints AI doesn't know — legacy system interface validation, data quality implications for migration architecture, Quebec Law 25 and OSFI data handling requirements, and the operational capability of the client's team. I document every significant decision with complete consequences — what the design enables, what it constrains, and what future architects must preserve. That documentation is part of my professional obligation to the engagement, not an afterthought."

Knowledge Check

A client's VP Technology asks in an SA engagement conversation: "We've had SAs use AI to generate architecture diagrams that looked great but missed key constraints — we ended up with integration designs that didn't account for our legacy system's limitations. How do you approach AI in your architecture work?" Which response positions you most effectively?

This response does three things that matter in this conversation: it directly names the failure mode the client experienced ("pattern answers that don't know your specific constraints"), it describes a specific practice that addresses it ("constraint layer"), and it names the deliverable that carries the knowledge forward ("Architecture Decision Record with complete consequences"). The client's concern about AI-generated diagrams that missed constraints is a legitimate concern about a real SA failure mode. This response shows you understand why that failure happens and have a specific practice to prevent it — not just generic reassurance that you're "careful." That specificity is what earns senior rate conversations.

Option 3 is the strongest response because it addresses the specific failure mode the client named with a specific practice. Option 1 positions AI as peripheral to your work — which undersells the value while still raising the concern about constraints. Option 2 says "careful review" without describing what that means — which is exactly the vague reassurance that won't distinguish you. Option 4 describes validating against documentation, which is part of good practice but doesn't address the core concern: some constraints aren't in documentation, they're discovered through direct engagement with the system and the team. Name the problem, name your specific response, name the deliverable that carries it forward.

4

Your SA capability readiness check

Answer based on what you can do and are doing today — not what you intend to do after this pathway.

I use AI for options generation — generating viable patterns with tradeoffs rather than asking for the best answer — and apply a specific constraint layer from the engagement context before selecting among them.

I write Architecture Decision Records with complete consequences — what the design enables, what it constrains, and what future architects must preserve — for every significant decision on an engagement.

I discover legacy system constraints directly before committing to integration designs — interface capabilities, throughput limits, message format requirements — rather than assuming from documentation.

I challenge my own architecture designs with an explicit AI review before presenting — asking for single points of failure, NFR gaps, and what a critical reviewer would raise — and address the findings proactively.

I produce audience-adapted architecture communication — different versions for CTO, development team, PM, and steering committee — verifying accuracy and consistency across all versions before distributing.

I can describe my AI-augmented SA practice specifically in a senior engagement conversation — including what AI generates, what constraint layer I apply, and why ADR documentation is part of my professional obligation to the engagement.

0/6

5

SA Accelerator — pathway complete

The insurance IT solution architecture market is changing. Clients have seen AI-generated architecture diagrams that looked complete and missed critical constraints. They've experienced the cost of discovering data migration architecture flaws during go-live. They're becoming sophisticated about what AI can and can't do in architecture work — and they're increasingly willing to pay a premium for SAs who understand the difference and can describe their practice specifically.

What this pathway builds is the professional standard for AI-augmented SA work: systematic use of AI for pattern generation and documentation, rigorous constraint discovery for things AI can't know, complete decision documentation that survives the engagement, and communication that serves every audience correctly. These habits compound. The SA who applies them consistently across two or three engagements builds a track record that is distinctly different from one who uses AI casually and delivers architectures that look right.

The SA who doesn't do this

Two kinds of AI-using SAs are emerging. One uses AI to produce polished architecture diagrams faster and relies on the visual confidence of the output to substitute for constraint validation. The other uses AI to generate the option space faster and spends the recovered time on the constraint discovery, assumption testing, and decision documentation that no pattern can replace. The second category delivers architectures that survive implementation. The first creates expensive surprises. This pathway is the difference.

🎓

SA Accelerator — Complete

Five modules. Constraint-first. Guidewire implementation and enterprise architecture contexts. You have an AI-augmented SA practice that is systematic, defensible, and senior-market-ready for insurance IT engagements.

Review Pathway Your Icon Profile Explore Full Membership →