Toronto · Principal-led · By referral and introduction
RKade is a boutique practice for boards, CEOs, and CIOs who need senior judgement inside the room: fractional CTO leadership, enterprise architecture and transformation roadmaps, and AI governance built to withstand regulatory scrutiny. No benches of juniors. No slideware. The principal does the work.
The moment you're in
Boards want an AI strategy this quarter. Regulators want evidence of control. The transformation programme is in its third year and its second reset. And somewhere in the middle of it, there is a leadership gap no job posting has filled. This is the work RKade is engaged for.
Agentic AI is arriving in your workflows faster than your control environment can absorb it. The question is no longer whether to adopt — it is whether you can demonstrate, to a supervisor, exactly how these systems are governed, gated, and audited.
A CTO departure, a CIO stretched across two mandates, a programme that has outgrown its leadership. Institutions lose quarters to interim drift. An embedded executive restores decision-making in weeks, not after a nine-month search.
Target states drawn for the presentation rather than the institution. Architecture decisions deferred until they become expensive. What is missing is rarely analysis — it is a principal in the room with the standing to make the call and own it.
01 — Services
Every engagement is led personally by the principal, scoped tightly, and documented to the standard your risk, audit, and regulatory functions will demand.
RKade embeds at CTO or CIO level — attending your executive table, owning decisions, managing vendors and teams — for the duration of a transition, a mandate, or a programme that cannot wait for a permanent hire.
Typically engaged by: CEOs and boards during executive transitions; CIOs carrying a dual mandate; programme sponsors whose initiative has outgrown its leadership.
Enterprise architecture fails when it describes a destination without a credible path. RKade builds target-state architectures and sequenced roadmaps grounded in your regulatory obligations, your legacy reality, and your funding envelope — then stays to defend them in the rooms where they are decided.
Typically engaged by: CIOs and chief architects resetting a stalled transformation; boards seeking an independent view before committing material capital.
The gap between an AI pilot and an AI estate a supervisor will accept is a control environment: how models are accessed, how agents are authorised, how every action is logged and reversible. RKade designs and charters that environment for institutions operating under real supervisory expectations.
Typically engaged by: CIOs and chief risk officers scaling AI beyond pilots; model risk teams facing supervisory questions they cannot yet answer with evidence.
02 — Approach
The same discipline whether the engagement runs six weeks or eighteen months. You always know which movement you're in, what it produces, and what decision it sets up.
Structured listening and evidence-gathering: leadership interviews, portfolio and architecture review, regulatory posture. Findings stated plainly, including the ones that are uncomfortable.
The target state and the path to it — options costed, trade-offs named, a recommendation made. Not three options and no view: a position, defended with the evidence from the assessment.
The decision machinery is stood up: Architecture Review Board, principles, decision records, control mappings. Governance designed to accelerate good decisions, not to slow all of them.
Execution under the governance just built — embedded leadership, milestone accountability, and reporting a board can rely on. The engagement ends with a documented handover, not a dependency.
03 — Engagement models
RKade takes a small number of engagements at a time, and conflicts are checked before any discussion of scope. Terms are agreed in conversation — this practice does not publish rate cards.
A bounded, senior assessment of a defined problem: an AI governance gap, a transformation that has lost its thread, a target state that needs an independent verdict before capital is committed. You receive findings you can act on and a recommendation you can defend.
Best for: a specific question that needs an authoritative answer, quickly and discreetly.
The principal sits in the seat — executive table, decision rights, team and vendor accountability — through a transition, vacancy, or mandate. Part-time in hours, full-weight in the room. Includes structured support for recruiting and onboarding the permanent successor.
Best for: leadership gaps that are costing the institution quarters, not weeks.
A retained relationship for executives who want a principal-grade second mind: monthly working sessions, review of material decisions before they are taken, and a direct line when something cannot wait for the calendar.
Best for: CIOs and CTOs who have the team, but want seasoned challenge at the decision point.
Chartering and standing up the machinery of technology governance: an Architecture Review Board with real decision rights, an AI governance committee aligned to supervisory expectations, principles and decision records that give both their authority.
Best for: institutions whose governance exists on paper but not yet in practice.
On fees: engagements are scoped and priced individually, after a confidential conversation about the problem — not before it. If RKade is not the right firm for the work, you will be told so in that first conversation, and pointed to who is.
Not sure where this fits?
Answer as your situation stands today. The recommendation is the engagement model RKade would propose in a first conversation — nothing is stored or sent.
Recommended starting point
04 — Selected engagements
Client names are never published. What follows is the shape of recent engagements, with outcomes stated as the institutions themselves characterised them.
Engaged as agentic AI use cases multiplied beyond the model risk team's line of sight. RKade designed the LLM gateway pattern and guardrail framework, mapped controls to OSFI E-21, and chartered the governance committee. The institution moved from case-by-case exceptions to a single, evidenced path from pilot to production.
A multi-year core modernisation had consumed two resets and the executive team's confidence. A diagnostic sprint separated the programme's genuine constraints from its accumulated habits; the rebuilt target state and sequenced roadmap were adopted by the board, and the principal stayed embedded through the first two delivery quarters.
Architecture governance existed as a review meeting that projects routed around. RKade rechartered the board with explicit decision rights, a principles set short enough to be used, and decision records that made rulings stick. Within two quarters, major initiatives were coming to the board early — by choice, because it made delivery faster.
References are available under NDA once a genuine engagement discussion is under way. Sectors and engagement shapes above are representative of the practice's work; identifying details are withheld as a matter of policy, not presentation.
05 — About the practice
RKade is the practice of a single principal: a technology executive who has held VP and CTO mandates inside regulated institutions, led architecture at enterprise scale, and spent years on the hyperscaler side of the table — running cloud and AI practices for the providers your institution buys from. That combination is the firm's entire proposition: someone who has sat in your seat, and in the vendor's.
The practice is kept small by design. A limited number of concurrent engagements, no hand-offs to junior staff, and a standing rule that if the principal cannot do the work personally, the engagement is declined. What you are buying is a specific person's judgement, and you get exactly that.
Toronto-based, working across North America — on-site where the work demands it, which at this level it often does.
06 — Insights
Occasional, substantive writing on the questions clients are actually asking. No content calendar, no listicles.
Agents that can act — spend, write, approve — change the evidentiary burden on your control environment. What a defensible governance model looks like, and the four artefacts you will be asked for first.
Practice noteRequest the note → Regulatory ReadinessThird-party and model risk expectations land hardest on institutions whose architecture decisions were never recorded. A field view of the gaps that surface in supervisory conversations, and how to close them before the letter arrives.
Practice noteRequest the note → Architecture GovernanceMost ARBs fail the same way: too late in the process, too broad in scope, no decision rights. The charter, membership, and cadence that make a board a place projects want to come to early.
Practice noteRequest the note →07 — Questions
08 — Contact
Write two or three sentences about the situation. You will hear back from the principal directly — typically within one business day — with either a time to talk or an honest view that RKade is not the right firm.
Your note is with the principal. Expect a personal reply within one business day, either proposing a time to talk or giving you a straight steer if RKade isn't the right firm for this.