proxem.ai  ·  /solutions03 services
Services

What we do. And
deliberately, what we don't.

Three defined services. Fixed scope. Fixed partner. Delivered by a team of no more than four people, led by the partner who quoted the work.

Service A

AI Strategy
& Portfolio Review.

Commissioned by the CEO or board. A clinical review of every AI initiative against commercial merit, regulatory exposure, and organisational capacity to execute.

Typical buyer: CEO, CSO, board audit committee
Deliverables
A.01
Portfolio census

Every AI project, pilot, and vendor contract catalogued with owner, spend-to-date, and stated outcome.

A.02
Commercial merit scoring

Each initiative rated on cost base attacked, time to value, and defensibility of the resulting position.

A.03
Regulatory exposure map

AI Act risk classification, GDPR Article 22 implications, and sector rules per initiative.

A.04
Capability-to-execute audit

Honest assessment of data, talent, and organisational readiness, against what each initiative requires.

A.05
Kill list

A named set of initiatives recommended for decommission, with reasoning and a transition plan.

A.06
Twelve-month roadmap

What to fund, in what sequence, under what governance. Explicit about what we refuse to recommend.

Process

How the engagement runs.

Step 01 · Week 1–2
Scope & access

Partner-led kickoff. Access to finance, risk, IT, and business owners established under NDA. Working documents exchanged.

Step 02 · Week 3–5
Evidence collection

Structured interviews, contract review, and on-site observation. Findings logged against a live register, visible to the sponsor.

Step 03 · Week 6–8
Synthesis

Hypothesis tested against CFO, CIO, and general counsel. Dissent recorded. Recommendations adjusted or defended.

Step 04 · Week 9–10
Delivery

Board-session read-out. Written report with evidence annexes. Thirty-day follow-up to review adoption.

Service B

Custom
Integration.

Your existing systems were built over years. We design and deploy AI integrations that work within that architecture — RAG pipelines, API orchestration, and data infrastructure that embeds cleanly without disrupting what already works.

Typical buyer: CTO, Chief Data Officer, VP Engineering
Deliverables
B.01
Integration architecture

The technical blueprint for embedding AI into your systems. Data flows, API contracts, and infrastructure requirements agreed before a line of code is written.

B.02
RAG pipeline build

Retrieval-augmented generation connecting your proprietary data to a language model. Built on your infrastructure, under your data governance.

B.03
API orchestration layer

The middleware that connects AI components to your core systems — ERP, CRM, data warehouse — with defined contracts and fallback behaviour.

B.04
Data pipeline engineering

Ingestion, transformation, and quality controls that prepare your operational data for AI consumption. Monitored from day one.

B.05
Integration testing & QA

Systematic validation across edge cases, load conditions, and failure modes. Delivered as a reusable test suite your team can run independently.

B.06
Handover & documentation

Complete technical documentation and structured handover to your team. Every component is client-owned on day one.

Process

How the engagement runs.

Step 01 · Week 1–2
Discovery & scoping

Current systems mapped, integration constraints identified, technical specification agreed. No build begins without a signed-off spec.

Step 02 · Week 3–8
Build & integrate

Iterative development with weekly demos on a shared staging environment. Scope changes managed transparently.

Step 03 · Week 9–12
Testing & hardening

Integration tested against your live environment. Failure modes documented and handled before any production release.

Step 04 · Week 13–14
Deployment & handover

Production release with rollback plan. Structured handover and 30-day stabilisation support included.

Service C

Agentic
Workflows.

The step beyond task automation. We design and build autonomous agents that manage complex, multi-step processes — operating continuously, reliably, and within governance boundaries your organisation can audit and control.

Typical buyer: COO, Head of Operations, CIO
Deliverables
C.01
Process audit & agent design

The target process mapped end-to-end, decomposed into agent-compatible steps, with decision boundaries and escalation logic defined.

C.02
Agent build

The autonomous system engineered, tested in isolation, and validated against defined performance criteria before any production exposure.

C.03
Governance wrapper

Audit logging, human override controls, and anomaly detection built into the agent architecture. Oversight is structural, not optional.

C.04
Pilot deployment

Controlled release on a defined slice of live traffic. Performance measured against the baseline from discovery. Issues resolved before full rollout.

C.05
Scale plan

A sequenced plan for extending the agent across full process scope, with risk assessment and a named kill-switch condition at each stage.

C.06
Operational handover

Runbooks, incident response procedures, and a named escalation path. Your team operates the system from go-live.

Process

How the engagement runs.

Step 01 · Week 1–3
Process audit

Target process mapped, baseline metrics established, agent scope agreed. Includes a formal assessment of whether an agent is the right solution.

Step 02 · Week 4–7
Architecture & build

Agent logic, toolset, and governance controls designed with your technical and operational leads. Built to specification.

Step 03 · Week 8–12
Pilot & adjust

Agent deployed on contained scope. Behaviour monitored daily against agreed metrics. Adjustments made before full rollout.

Step 04 · Week 13–16
Production & handover

Full deployment with governance framework active. Documentation, training, and a 60-day post-launch review.

/ engageresponse < 48h

The briefing comes
before the engagement.

We do not quote work without a 90-minute conversation with the people accountable. It is the only way to know whether we should take the engagement.