Three ways to engage. The methodology stays the same.

When the problem is clear but the starting point is not, we embed with your team and run a process discovery before we build anything. The approach is what you would expect from a top-tier process-assessment engagement (Centric, Gartner, Blue Prism patterns).

Map workflows as they actually run. We walk the floor with the people doing the work, not with the documentation.

Identify bottlenecks: rework loops, manual handoffs, approval delays, error-prone steps, tasks that have clear rules but no automation.

Score each process on four dimensions: volume, complexity, business impact, automation feasibility. The processes with high volume, low complexity, and high error rates go first.

Produce a prioritized roadmap. The top two or three candidates become Phase 1 builds, each running the DMAIC lifecycle above.

Typical timeline: 2 to 3 weeks for the discovery, then 8 to 14 weeks per build.

When you have AI in production or under evaluation and you want to know whether the architecture, the governance, and the risk posture hold up, we run an assessment against the frameworks that already apply to your context.

Inventory. What AI systems exist. What they do. Who owns them. Including third-party AI embedded inside vendor products.

Governance assessment. Are roles, escalation paths, and decision rights documented and operational? Is the human-in-the-loop model tiered deliberately or by default?

Risk assessment against the standards that apply: NIST AI RMF, ISO 42001, COSO adapted for AI, or the IIA AI Audit Framework. We use whichever your organization already operates under, or recommend the best fit if you have not adopted one.

Gap analysis. Where the current architecture falls short. Which gaps carry the most risk. Which are cheapest to close.

Roadmap. Prioritized recommendations with effort bands and a suggested sequence.

Typical timeline: 3 to 6 weeks.

The estimate comes back within 48 hours of the call.

Every estimate has two scenarios. Scenario A: if the working assumptions hold. Scenario B: if a key dependency does not hold. The dependency is named so neither of us is surprised later.

The estimate separates Phase 1 (discovery) from Phase 2+ (build). Some clients commission only Phase 1 first, with a decision gate after.

If you need a rough size before a call to know whether this fits your budget, we can give you a band once we have heard the shape of the work.

Every engagement contract sets out: scope (what is in, what is out, what sits on the production gap list), deliverables by phase with named acceptance criteria, timeline with two-scenario framing, pricing (flat-fee or fixed-fee-per-phase), data handling provisions, the 30-day stabilization window, and termination clauses that protect both sides.

No innovation fees. No value-share structures. No retainer that grows quietly.

Every effort estimate ships with two scenarios. Every status report uses the same numbers across audiences, framed differently for context. If a program is two weeks behind, the report says two weeks behind, not "tracking to plan with minor adjustments."

Recommendations are explicit, but they remain recommendations. The decision belongs to you. We do not soften the trade-off to make our preferred answer easier to accept.

New scope gets its own envelope. "While you are in there" is a signal to write a change request, not to silently absorb the work.