What’s included
What’s included
AI threat modelling
Prompt injection, data exfiltration, tool abuse and model supply chain risks mapped to your architecture.
Built with
- Amazon Bedrock Guardrails
- AWS IAM
- AWS PrivateLink
- AWS CloudTrail
Why it matters
Why it matters.
AI security questionnaires now gate deals.
A clear control story turns a blocker into a differentiator.
Prompt injection, data leakage and over-permissioned tools are new risk classes your existing controls were not designed for.
Regulators and auditors are catching up fast. Building the evidence trail now is far cheaper than retrofitting it.
What’s included · how it runs
Assess. Remediate. Evidence.
Step 01
Assess
Threat model the AI workloads and review current controls and data flows.
What’s included
AI threat modelling
Prompt injection, data exfiltration, tool abuse and model supply chain risks mapped to your architecture.
Step 02
Remediate
Implement guardrails, data protection and access boundaries.
What’s included
Data protection
Classification, redaction, retention, residency and private connectivity for model traffic.
Guardrails and controls
Bedrock Guardrails, input and output filtering, least-privilege tool access and human-in-the-loop where it matters.
Step 03
Evidence
Wire up logging and produce the control documentation buyers and auditors ask for.
What’s included
Compliance evidence
Control mapping, logging and artefacts aligned to SOC 2, ISO 27001 and your customer questionnaires.
What you walk away with
- AI threat model and risk register
- Guardrails and data protection controls
- Audit-ready control mapping
- Reusable answers for security questionnaires
Case studies
Customers we’ve delivered this for.
Draft · pending HES approval
Case studyHuman eSources
Make the hiring call with AI.
How Human eSources turned nearly three decades of talent science into AI that explains every job match, live on Amazon Bedrock behind TruTalent, its new recruiting platform.
- Industry
- HR tech, talent assessment
- We built
- Talent Scout AI
- Runs on
- Amazon Bedrock
Keith Gaudreau
VP Product Strategy
NeuroMind
Draft · pending NeuroMind approvalCase studyNeuroMind
A loan platform built for bank-grade security.
How NeuroMind modernised the loan management platform it runs for a major bank on AWS: secure by design, PCI-DSS aligned and quicker to release.
Read the case studyCase study[Client]
[The promise] [in 2 lines].
[How the client got the outcome: what Assembly built and where it runs. One sentence.]
Talk to an engineer[Client logo][First Last][Role]
Related beyond Gen AI