Services
Readiness, pilots, governance, and a seat at your table.
Three scoped engagements and one standing relationship. Each is fixed-fee, scoped and itemized before you commit, and each produces something your team owns and can operate without us.
Where a new client starts
Four steps, and you can stop at any one of them
Most organizations begin with the readiness audit and add the rest as their AI footprint grows. Each step stands on its own.
AI Readiness Audit
Can this organization support AI safely, across its people, its processes, and its data, what is already running without anyone approving it, and which two or three opportunities pay back first?
- Organization readiness: leadership alignment, decision rights, budget ownership, risk appetite.
- People readiness: current skills and usage, role-level impact, training needs, where adoption will meet resistance.
- Process readiness: which workflows are documented, repeatable, and high volume enough to be worth redesigning.
- Data readiness: where data lives, its quality and access paths, ownership, what must be cleaned or governed first.
- Shadow AI discovery: the tools your people already use on company data without approval, plus the AI features switched on inside software you bought for something else.
- Regulatory exposure: the countries, regions, and states you operate, sell, hire, and hold data in, mapped to the rules that follow. We ask, we do not assume, and you keep the workbook.
- AI opportunity register: candidate use cases scored on reach, impact, confidence, and effort, with an ROI estimate.
- Quick-win shortlist: the two or three opportunities that pay back fastest at acceptable risk.
Independent Pilot Design & Oversight
Is this idea worth building, and did the thing that got built do what it was supposed to do?
- Pilot design: the problem stated plainly, a measured baseline, acceptance criteria agreed before anyone writes code.
- Boundaries set first: which data the system may touch, which permissions it gets, where a human has to sign off.
- Build oversight: partner selection, requirements, progress cadence, and acceptance testing against the criteria.
- A go, adjust, or stop decision at the end, argued from the measurement rather than from enthusiasm.
- Independent by structure. We take no build fee on the work we oversee, so the party checking the result is never the party paid to ship it.
Governance, Compliance & Security
Who approves what, can you prove the controls work, and would the security posture survive a customer, a prime contractor, or a regulator looking at it?
- AI use policy, decision rights, approval workflow, and lifecycle controls for models, prompts, and agents.
- Obligations mapped to controls, owners, and evidence against NIST AI RMF, ISO/IEC 42001, the EU AI Act, SOC 2, CMMC, or your sector equivalent.
- Security review: data flows, model access paths, least-privilege permissions for agents, audit trails, and what happens when a system acts outside its scope.
- Vendor and third-party AI due diligence, with questionnaire and flow-down clause language.
- Gap audit and conformance testing: whether your own policies are followed in practice, including shadow AI and vendor-embedded AI.
- An evidence pack you can hand to a customer, an insurer, a prime, or an auditor. Repeatable, so new builds are checked as they ship.
Scale, Continuous Oversight & IT and AI Strategy
Who takes this from one working pilot to an operation, keeps the controls honest as it grows, and tells you what to build next?
- A fractional Chief AI Officer seat at your table, month to month, without an executive salary.
- Scale support: moving a proven pilot into production across teams, with the integration, platform, and cost decisions that come with it.
- Continuous oversight: policy upkeep, control re-testing, periodic security checks, board and customer reporting.
- IT and AI strategy: roadmap, build-versus-buy calls, vendor selection and negotiation, budget planning.
- Market watch. We filter the noise and bring you the few developments that change your decisions.
- Available with or without forward-deployed engineers who sit with your team and build automations and AI releases continuously.
Build work runs through vetted partners and forward-deployed engineers, with HEED SAI holding the contract. We take no build fee on work we also audit, so the party checking the control is never the party paid to ship it.
Also available
Training and executive education
For leadership
What is possible, what it costs, what it risks, and what a responsible decision looks like. Regulatory framing for GDPR, the EU AI Act, and the US state rules that already apply to you, explained without jargon.
For teams
Foundational through industry-specific sessions. Real concerns and real risks addressed in plain language, so hesitation turns into a team that knows where the guardrails are.
What you receive
Deliverables you own
AI Readiness Assessment
Organization, people, process, and data readiness with scores, gaps, and owners.
AI Opportunity Register
Scored use cases with ROI estimates, a shadow AI inventory, and a quick-win shortlist.
Pilot Design & Acceptance Brief
Problem, measured baseline, acceptance criteria, data and permission boundaries, human sign-off points.
AI Governance Pack
Policy, standards, council charter, and the approval workflow your team runs.
Compliance & Security Control Matrix
Obligations mapped to controls, owners, and the evidence that proves them.
Gap & Findings Report
Governance, compliance, and security gaps with severity, impact, owners, and a remediation roadmap.
Executive Readout
Findings, the decisions required of you, and the recommended sequence, in one session.
Every deliverable is produced for your review before it is final. Nothing is filed, submitted, or published to a third party, a regulator, or a customer without your written authorization.
New assignments by referral only
We have been humbled by the amount of work arriving through referrals from clients and colleagues. To protect the quality of what we deliver to the businesses already counting on us, HEED SAI is accepting new assignments by referral only for the moment.
If someone we work with pointed you here, write to info@heedsai.com and tell us who sent you. We read every message, and we will tell you honestly whether we have room.