The problem: Leadership can't honestly say where the organisation stands on data, skills, governance, and AI use cases.

Ideal for SMEs that know they should be doing something with AI but don't know where to start, or boards needing a defensible baseline.

What you get:

• Structured scoring across five dimensions: data, technology, skills, governance, and use-case potential
• Written report with a maturity score, top three risks, and prioritised next steps
• Findings session with the leadership team to agree the next move

The outcome: An honest baseline, a board-ready answer, and the next two or three moves named — not a glossy report that gets filed.

The problem: AI ambitions exist on a slide but there's no agreed sequence, no budget, and no view on which bets pay back first.

Ideal for SMEs that have completed an assessment and now need a credible plan to back investment decisions.

What you get:

• Use-case shortlist with value, feasibility, and dependency analysis
• Sequenced roadmap across with indicative budget envelopes and a build-vs-buy view
• Risk register and board-ready summary deck the sponsor can present without rework

The outcome: A plan the leadership team will fund, with decision gates that survive contact with reality.

The problem: AI is already in use across the business often without sign-off and no one has mapped it against the EU AI Act or GDPR.

Ideal for SMEs selling into regulated sectors or operating AI systems that may fall under prohibited or high-risk categories.

What you get:

• Inventory and risk classification of current and planned AI systems against the EU AI Act
• Gap analysis against documentation, transparency, and human-oversight obligations
• Board-level briefing on obligations, exposure, and penalties plus an acceptable-use policy for staff

The outcome: Documented evidence that your AI use is lawful and defensible to customers and regulators.

The problem: AI is too important to ignore but too early-stage to justify a full-time hire.

Ideal for SMEs in the €5–50M revenue band that need senior AI judgement at the leadership table without a permanent hire.

What you get

• Named senior Talos partner acting as part-time Head of AI, with a seat at the leadership table
• Board-level reporting on AI initiatives, spend, and outcomes
• Hiring, mentoring, and capability-building support — plus first call on Talos delivery capacity

The outcome: AI decisions stop being made by accident. Strategic ownership without the cost of a full-time CAIO.

The problem: AI initiatives keep stalling on the same answer: "the data isn't ready."

Ideal for SMEs whose data lives across spreadsheets, line-of-business systems, and inboxes and who need it governed before AI sits on top of it.

What you get

• End-to-end review of data collection, ingestion, storage, and access controls
• Data quality assessment and prioritised remediation plan
• Reference architecture and governance model covering ownership, retention, and access

The outcome: Data that AI systems can actually be built on, with the ownership and controls to keep it that way.

The problem: Multi-step, cross-system processes are still being done by people copying data between SaaS tools. 

Ideal for SMEs with repetitive, multi-system processes — procurement, claims, onboarding, scheduling — where the value is in orchestration, not just the language model.

What you get

• Agent architecture design with clear scope, tool boundaries, and human-in-the-loop checkpoints
• Tool and API integration with existing SaaS systems, plus an evaluation framework and guardrails
• Production deployment with observability, rollback paths, and an operations playbook

The outcome: A process that used to require people now runs itself, safely, with the right.

The problem: Institutional knowledge is locked in SharePoint, Drive, and old email archives. 

Ideal for SMEs whose knowledge is scattered across disparate systems and who want staff to ask plain-English questions and get cited answers.

What you get

• RAG architecture with citation and audit trail, integrated with Microsoft 365, Google Workspace, or other source systems
• Permissions model that respects existing document-level access
• Evaluation suite covering accuracy, hallucination rate, and refusal behaviour

The problem: Everyone on the leadership team has a different picture of where AI fits so every conversation ends up too broad to act on.

Ideal for SMEs that know AI matters but can't name the three highest-value places to apply it.

What you get:

• Pre-workshop interviews with department leads followed by a facilitated opportunity-mapping session
• Prioritised opportunity backlog scored on value, feasibility, and risk
• One-page summary ready to brief the board

The outcome: The leadership team agreeing in writing on the three places to start, the three to defer, and the three to leave alone.

The problem: Copilot, ChatGPT, or Claude licences have been bought. Usage is concentrated in two or three enthusiasts.

Ideal for SMEs that have licensed AI tools but aren't seeing adoption, or any team building AI fluency across the organisation.

What you get:

• Role-specific training tracks for executives, knowledge workers, sales, operations, and developers
• Internal prompt library tailored to the organisation's actual work, plus live workshops and recorded micro-lessons
• Optional champions programme and usage reporting to embed capability without ongoing reliance on Talos

The outcome: Licences that are actually used, daily, by people who can explain what they're doing and why.