Architecture review

Before committing significant budget to OculiX — or when an existing setup feels brittle and you want to know why — a structured architecture review gives you a second pair of expert eyes on your visual automation strategy. The output is a written report, not slides, and is yours to keep regardless of what you decide next.

Who it’s for

Teams about to commit budget

You’re evaluating OculiX (or comparing it to alternatives) for a meaningful deployment. Before signing off on the architecture, you want an independent technical opinion.

Teams with brittle existing setups

Your OculiX (or other RPA) deployment works most of the time, but breaks often enough to be a nuisance. You suspect there’s a structural reason but you’re inside the problem and can’t see clearly.

Teams considering migration

You’re on UiPath / Automation Anywhere / Blue Prism / TagUI and weighing whether OculiX is the right destination. You want to understand the structural implications before committing to the Migration engagement.

Teams pushing scale boundaries

You’ve made OculiX work for one team, now multiple teams want in. Scaling from 10 scripts to 1,000 surfaces architectural questions that didn’t exist before.

Teams in regulated sectors

Banking, defense, healthcare, government. Your CISO / risk team asks pointed questions about OculiX and you need a credible answer beyond “the docs say it’s secure”.

Teams hitting performance walls

Tests that ran in 30 min now take 4 hours. Memory grows unbounded. CI queues back up. You suspect tuning, you suspect architecture, you want to know which.

What’s covered

Stack assessment

Inventory of your OculiX deployment: versions, modules in use, scripts catalog, integration points, scheduling infrastructure, monitoring setup. Catalog with severity labels for known issues.

Pattern review

Review of representative scripts and helpers for patterns that scale vs patterns that don’t. Common anti-patterns flagged with recommended refactor priority.

Scaling analysis

Your current load, projected growth, bottlenecks in the path. Where parallel execution helps and where it breaks. JVM tuning, process isolation, CI orchestration.

Security posture

Application of OculiX security architecture to your specific deployment. Air-gap considerations, secrets management, audit trail (MCP module), supply chain (SBOM consumption).

CI/CD integration

How OculiX fits in your pipeline today. Where it adds friction, where it should integrate more deeply. Tradeoffs of running scripts in CI vs scheduled jobs.

Monitoring & observability

What you observe today, what you should observe, what tooling fits. Failure mode visibility, alerting, audit retention, post-mortem readiness.

How it works

1. Scoping call (free, 30 min)

We understand your context, the questions you want answered, and the depth that fits your situation. Most reviews fall into one of the 6 buckets above, but the actual focus is shaped by what matters most to you. If a particular concern dominates (e.g., “we’re stuck on a 90-min CI suite, can we get to 30?”), the review focuses there.

No commitment from this call — we’ll either propose to proceed, suggest a different engagement (e.g., training instead), or decline if architecture review isn’t the right fit.

2. Information gathering (3-5 days)

You provide, under NDA if needed:

Read access to your OculiX scripts repo (or a representative sample)
Your CI configuration files
A walkthrough call (60-90 min) with the team that operates OculiX day-to-day
Any relevant ops docs, runbooks, post-mortems, monitoring screenshots
The questions you want answered, ranked by priority

We don’t need access to production systems — just enough material to understand the architecture in depth.

3. Analysis (5-8 days)

We work through the material, reproduce relevant patterns in a sandbox where possible, benchmark where bottlenecks are suspected, draft the report. We may ask 1-2 short clarification questions during this phase but mostly work independently.

4. Written report (delivered end of Week 2)

A structured PDF document, typically 15-30 pages, covering:

Executive summary (1 page) — top 5 findings, top 5 recommendations, in plain language for non-engineers
Current state — your deployment described in our words, so you know we understood correctly
Gaps & risks — what’s structurally fragile, what’s likely to fail under specific scenarios
Recommendations — prioritized by ROI (impact / effort), with concrete next steps
Implementation roadmap — what to tackle first, second, third, and how long each is likely to take
Open questions — things we couldn’t determine from the material that need a decision on your side

The report is yours. You can share it internally, use it in budget conversations, use it as a baseline for a vendor selection process. We don’t retain rights over it beyond reference for our records.

5. Follow-up call (45 min, included)

A live walkthrough of the report with the people on your side who’ll act on it. Q&A, clarifications, prioritization adjustments based on internal constraints we didn’t know about. No surprise findings revealed in the call — everything substantive is already in the report.

After this call, the engagement is complete. If you want us to implement the recommendations, that’s a separate engagement (Custom development or Production support depending on scope).

Examples (anonymized)

Bank — 40% CI suite speedup

A bank running 50,000 OculiX executions per day across 12 CI workers asked: “why is our suite 90 minutes?”. The review identified: JVM cold-start dominating, no OpenCV pre-load, suboptimal parallelism on the workers. After they implemented the 3 recommendations: 54 minutes. The review fee paid back in CI minutes saved within 6 weeks.

Insurance — security pre-audit

A French insurance group preparing for an internal security audit got a review of their OculiX deployment for the audit committee. Report covered air-gap readiness, audit journal usage, secrets handling, SBOM consumption, CVE response. Audit passed with zero findings on the OculiX component. The report was attached as evidence.

Defense contractor — air-gap migration

A defense contractor moving from a connected environment to an air-gapped one. Review covered build-from-source on air-gap, dependency mirroring, signed-release verification, audit trail compliance. Migration succeeded in 6 weeks with zero air-gap leaks. The report became the team’s internal runbook.

Manufacturing — scaling from 10 to 200 scripts

A factory engineering team grew their OculiX usage from 10 scripts to 200 over a year and was hitting maintenance walls. Review identified: helper sprawl, image-anchor naming chaos, no shared library across teams. Recommendations: consolidate helpers into a shared Java module, naming convention, central image library with versioning. Reorg completed in a quarter, maintenance burden dropped 60%.

Frequently asked questions

How long is the scoping call?

30 minutes by default, sometimes 45. We don’t bill for it, and we don’t push to convert it into an engagement — about 30% of scoping calls conclude “your situation is actually fine, no review needed” or “training would help more”.

How deep does the report go?

It’s an architecture review, not a code audit. We won’t read every line of every script. We read enough to identify structural patterns, then describe them at the right level for a technical decision-maker. Typical report length: 15-30 pages of substantive content (not padding).

Will you implement the recommendations?

Not in this engagement. The review is strictly the assessment + report + follow-up call. If you want implementation, we’ll scope it as a separate engagement (Custom development, Production support, or a structured implementation project). The review is standalone by design — you can take the report and implement internally, or with another partner, or not at all.

Are findings confidential?

Yes. We sign an NDA before access if you require one (standard NDA template provided, mutually adjustable). The report is delivered only to you. We don’t publish anonymized case studies without your written approval.

What if the review finds OculiX isn’t the right tool for our case?

We’ll say so in the report. We’ve recommended against OculiX adoption twice in past reviews where the use case was better served by another approach (one was a web-only DOM-heavy app where Playwright fit better; one was a mainframe automation case where green-screen drivers were more appropriate). Honest assessment is the value, not “always say yes to OculiX”.

Do you cover non-OculiX tools in the review?

The review is OculiX-centered. We compare OculiX to alternatives factually when relevant (e.g., “for this case, OculiX + Selenium combined would fit better than OculiX alone”). We don’t do generic RPA strategy consulting — there are firms that specialize in that.

How much does it cost?

Fixed price per engagement, scoped on the call. The fee correlates roughly with the depth and breadth of the analysis. Most reviews land in a comparable range — predictable for budget purposes.

Can we have multiple reviewers from your side?

By default, the maintainer (Julien Mer) leads the review. For very specific topics (e.g., security architecture for a defense audit), we can bring in a domain specialist as co-reviewer. Discussed during scoping.

How does this compare to hiring a generic consultancy?

A generic consultancy will give you broad RPA-strategy advice. We give you OculiX-specific architectural advice from the people who maintain the codebase. Different scope, different depth, different fit. Some teams hire both for complementary perspectives.

Get in touch

Request a free scoping call Tell us in a few lines what you'd like reviewed. We'll set up a 30-min call within 48h and propose an engagement or a different path if it fits better.

Read the security architecture first If your review interest is security-driven, the public Security page covers most of the standard concerns.

🦎