AI Analyst

About the project (description, duration, stage)

Join Neurons Lab as the AI Analyst on a flagship engagement with a European private investment group — a holding company with a C-level executive team, an investment/portfolio function and an affiliated family office.

The programme builds one private, access-scoped context layer over the group's data and then AI skills and agents on top of it — first for the executive team (6–10 people holding ~90% of group priorities), then for every employee. Two loops run on the same layer: alignment — are we doing the right things (strategy, OKRs, drift by team and by person) — and efficiency — are we doing things right (a process miner reads real workflows from the digital footprint, optimizer agents then implement the fixes).

Your half of the programme is the part that only a human can do. Phase 3 (Distill) is yours: sitting with each executive and getting what is in their head into the layer — group strategy and OKRs from the Chief of Staff, investment policy and portfolio base from the CIO, reporting standards and operating processes from the CFO and COO. And in the efficiency loop, you turn raw mining output into a written optimisation report per team: automate, reorganise, or leave alone — with the financial case attached.

Four phases — Capture → Connect → Distill → Build — over roughly eight to ten two-week sprints, opening with a fixed-fee two-week Sprint 0 readiness pass.

Stage: pre-contract / design-partner negotiation.

Duration: multi-phase, ~4–5 months to the executive pilot in production, then rollout.

Reporting: CTO and CEO are in the room at key points; you work day to day with the AI Architect (1.0 FTE) and a Data Engineer (0.5 FTE), and directly with the client's executive team.

Full-time role.

This is the most client-facing seat on the pod after the founders.

What you'll actually do (example tasks)

• Run executive distillation sessions — one-to-one with the Chief of Staff, CIO, CFO and COO — and turn each into a context pack: goals, OKRs, KPIs, investment policy, reporting standards, operating processes written down as usable text, not slides.

• Elicit and validate the business semantics of the ontology with stakeholders: what a "commitment", "decision", "priority", "portfolio update" actually mean in this group, and where definitions conflict between entities.

• Specify the agent skills per executive — scope, inputs, outputs, tone, acceptance criteria, escalation and human-in-the-loop boundaries — and write the evals that decide whether a skill is good enough to ship.

• Design the weekly alignment ritual in Slack: OKR-coached check-ins, drift detection, and the board master-report that assembles itself from the check-ins.

• Interpret process-mining output into a decision-ready report per team: where effort actually goes, what to automate, what to reorganise, what to leave alone — each with an ROI estimate and a recommended sequence.

• Build quick prototypes (no-code / low-code / prompt-level) to test a skill with an executive before engineering builds it properly.

• Own adoption: sit with the executives, watch them use it, find why they don't, and feed that back into the backlog every sprint.

• Measure payback after each automation ships and re-prioritise the next wave against it.

• Keep the written trail — decision records, requirement docs, runbooks — so the client's own team can eventually build the rest without us.

Skills

• Executive stakeholder management and workshop facilitation — can hold a room of C-level people and leave with something written down

• Process analysis and mapping: current-state documentation, process-as-is vs. process-as-written, workflow redesign

• Requirements engineering for AI systems: user stories, acceptance criteria, eval design rather than vague wish-lists

• ROI / business-case modelling and prioritisation under constraints

• OKR / goal-management fluency — enough to coach, not just record

• Hands-on with LLM tooling: prompting, no-code/low-code prototyping, agent builders, evaluating output quality critically

• Comfortable reading process-mining / usage data and reasoning about it quantitatively (SQL or spreadsheet-level analysis is enough)

• Exceptional written English — most of your output is prose someone else acts on

Knowledge

• Financial services / private equity operating context: investment policy, portfolio reporting, board and committee process, family-office structures — a strong plus

• AI governance basics in regulated environments: what to document, what needs a human, what needs an audit trail

• GDPR fundamentals as they apply to employee-generated data (mail, chat, meeting recordings) — including the politics of capture-by-default

• Awareness of ontologies / knowledge graphs — you don't build them, but you must be able to argue about definitions with the architect

Traits

• Strategic thinker who can also do the unglamorous documentation work

• Comfortable telling an executive their stated process isn't the one the data shows

• Technically curious and genuinely hands-on with AI tools, without pretending to be an engineer

• Bias to writing things down; allergic to unresolved ambiguity

Experience

• 4+ years in business analysis, management consulting, process improvement or AI/product analysis

• Demonstrated experience eliciting requirements from senior stakeholders and shipping against them

• Hands-on LLM / generative-AI implementation experience — prototypes you can show, not courses you attended

• Experience mapping and redesigning real business processes, ideally with mining or usage data rather than interviews alone

• Background in or with financial services / investment firms — strong plus

• Comfortable as the sole analyst on a small (2.5-FTE) delivery pod