Diagnostics
Start with evidence, not a build
A major platform or AI build is hard to scope before anyone understands the problem underneath it. Each MDA diagnostic is a fixed-scope, two-to-three-week engagement that gives a sponsor enough evidence to decide what to fund, sequence and control next.
Every diagnostic is led by Gaety, ends with an executive readout and closes with a clear next decision. The recommendations are yours to act on, with MDA or without.
Which one fits?
| Your situation | Start with |
|---|---|
| Reports disagree, sources are fragmented or modernisation priorities are unclear | Data Trust & Transformation Diagnostic |
| You have a specific AI or agentic workflow in mind and need to know if it can work responsibly | AI Use-Case & Data Readiness Assessment |
| An AI or automation tool is live, but it is unreliable, poorly controlled or not being used | AI Reliability & Adoption Diagnostic |
Not sure? Describe the situation and we will recommend the right starting point.
Data Trust & Transformation Diagnostic
2-3 weeks · Founder-led
For leaders facing untrusted reporting, fragmented source systems or unclear modernisation priorities, who need to know where to start before committing to a major platform build.
What we look at
- Up to six stakeholder interviews across business, data and technology
- A targeted review of architecture, representative data flows, reports, models, controls and operating practices
- One executive prioritisation workshop
What you receive
- A current-state assessment and a maturity and readiness heatmap
- The principal risks and dependencies
- Target-state principles and an opportunity map
- A prioritised 12-month roadmap with a recommended first work package
- An executive readout
What it is not: a penetration test, a legal or compliance certification, a full data-quality profiling exercise or a detailed physical solution design.
AI Use-Case & Data Readiness Assessment
2-3 weeks · Founder-led, with specialist review where needed
For sponsors with one concrete AI or agentic decision or workflow in mind, who need a credible go or no-go before investing in a pilot.
What we look at
- One defined use case, with sponsor and user interviews
- The source data and context the workflow would rely on
- Permissions, quality and lineage
- Retrieval options, evaluation needs and human-oversight requirements
What you receive
- A go, conditional-go or no-go verdict, with the evidence behind it
- A gap assessment and reference architecture
- A governed context design and recommended retrieval pattern
- An evaluation and control plan
- A sequenced pilot roadmap
What it is not: a promise of a production AI system, model accuracy or a business outcome. A pilot is scoped separately and may involve a verified specialist delivery team.
AI Reliability & Adoption Diagnostic
2-3 weeks · Founder-led, with specialist input where needed
For sponsors of an AI or automation solution that is already deployed but produces unreliable answers, falls short of control expectations or is not being used.
What we look at
- One deployed solution or workflow, with sponsor and user interviews
- The business use case, source data, context and retrieval, and prompts or configuration
- Permissions, lineage, evaluation and monitoring
- Workflow fit, usage and adoption, and available cost and performance evidence
What you receive
- A root-cause assessment
- A reliability and adoption scorecard
- Remediation options with technical and control recommendations
- A prioritised 30/60/90-day plan
- A clear recommendation to repair, redesign, replace, retire or scale, presented in an executive readout
What it is not: a rebuild, production model tuning, penetration testing, legal compliance certification or a guarantee of accuracy or adoption. Specialist testing and implementation are scoped separately.
How a diagnostic runs
Step 1
A focused conversation
We confirm the decision you face, the sponsor and what evidence exists.
Step 2
Scope and access
We agree the interviews, systems, documents and dates, so the timetable holds.
Step 3
Evidence
Interviews, reviews and working sessions over two to three weeks.
Step 4
Readout and next decision
Findings traced to evidence, with owners, sequencing and the next decision stated plainly.