Business and operating model design that starts where value is actually created or lost — the decisions, the decision rights and the handovers between them — then redesigns the processes that carry them out, on system evidence rather than workshop opinion.
of employees inside a new structure agree it will achieve its goals. 88% of their leaders do.
Bain & Company, 2026 · n≈1,000of organisations have fundamentally redesigned a workflow — the single strongest driver of EBIT impact from GenAI
McKinsey, 2025 · n=1,491of senior technology leaders say their operating model must change within 12–18 months
Deloitte, 2026 · n=662Australian labour productivity growth a year since 2015, against a 60-year average of 1.6%
Productivity Commission, 2026A structure tells you who reports to whom. An operating model tells you how work actually gets done — which capabilities have to be excellent, who decides what, where the handovers are, what information moves between them, which partners do what, and how performance is managed. Change the first without the second and you get a new chart sitting on top of old behaviour.
Roughly two-thirds of leaders have been through an operating model redesign in the past two years, and half expect another within two. Completion rates have climbed from 51% to 79% since 2014 — but completion is not the same as effect.
McKinsey, June 2025 · n=2,000
88% of leaders believe their new structure will achieve its goals. Only 36% of the employees working inside it agree, and just 22% say they received sufficient training, coaching or tools to adapt to the new way of working.
Bain & Company, January 2026
Nearly 75% of senior technology leaders say their operating model will need to change within 12–18 months, and 42% expect more than 40% of organisational processes to be automated or AI-enabled by 2028. Agents do not read org charts. They execute processes.
Deloitte, June 2026 · n=662
We use the Operating Model Canvas (Campbell, Gutierrez & Lancelott) as the spine, because it forces the conversation past structure — combined with capability-based design from the TOGAF standard, business architecture per BIZBOK, and the APQC Process Classification Framework as a common process taxonomy.
The work that has to be done to deliver the value proposition — end to end, across the functions that each own a piece of it.
Where does the value stream actually break?The people who do the work and how they are organised — roles, accountabilities, spans, and above all decision rights.
Who can actually decide, without escalating?Where the work is done, and the buildings, assets, sites and service centres it depends on — including what is centralised and what is not.
Is the shared service actually shared?The systems and data that support the work — and, just as importantly, the information that has to move across a handover for the next step to be possible.
What breaks at the handover?Which partners provide inputs to the work, what relationship exists with them, and what has quietly become a critical dependency without being governed as one.
Who owns the outcome, not the SOW?Planning, budgeting, performance management, risk, continuous improvement and people processes — the machinery that decides whether the new model survives its first bad quarter.
What behaviour do the KPIs actually reward?Capability maps and operating model choices are necessary. They are not sufficient. Value arrives through a sequence of steps, systems, decisions and handovers that turns an input into an outcome a customer or citizen actually experiences — and that sequence is almost never the one on the wall.
McKinsey tested 25 organisational attributes against GenAI's bottom-line effect. The redesign of workflows had the biggest effect on an organisation's ability to see EBIT impact. Only 21% of organisations had fundamentally redesigned any workflow. A year later, high performers were nearly three times as likely to have done so.
Australia has the definitive case study in the other direction. Robodebt automated an unlawful process at national scale and reached more than 500,000 people. The Royal Commission called it "a costly failure of public administration, in both human and economic terms." Automation does not fix a broken process. It industrialises it.
Illustrative of a typical order-to-cash baseline. On an engagement, both sides are generated from your own event logs.
The two most common failures in this work are diagnosing from opinion and declaring victory at go-live. The method is built to close both.
Process mining on your event logs, task mining where the work happens on the desktop, a capability heat map and a cost-to-serve view. We establish what the process is before anyone argues about what it should be.
Strategy translated into a capability model, then into the operating model choices that follow from it: what must be world class, what must simply be reliable, what should be bought rather than built.
All six dimensions designed together — processes, organisation and decision rights, locations and sourcing, information, suppliers, and the management system that will hold it in place.
Value streams redesigned across function boundaries. Fit-to-standard tested first. BPMN 2.0.2 and DMN 1.5 where decision logic is the real work; service blueprints where the customer is in the process.
Conformance checking against the redesigned model, simulation of throughput and capacity, and a costed benefit case for each redesigned process — before a line of configuration is written.
Role design, capability uplift, and the KPI and management-system changes that make the new way rational for the people doing it. Continuous conformance monitoring so the process cannot quietly revert.
Every enterprise system already records what actually happened, with timestamps. Process intelligence platforms reconstruct the real end-to-end process from those event logs — every variant, every rework loop, every wait state — and attach frequency and cost to each. It turns a workshop debate into an evidence review.
Discovers organisation-wide processes from event logs in your systems of record. Best for end-to-end flows — order-to-cash, procure-to-pay, hire-to-retire, permit-to-operate — where the inefficiency lives between functions rather than inside one.
Zooms in to the desktop, where the swivel-chair work lives: the copy-paste, the spreadsheet reconciliation, the workaround nobody documented. This is usually where the automation business case is actually found.
Drops the assumption that every event belongs to exactly one case. Orders, deliveries, invoices and payments are modelled as related objects — which removes the convergence and divergence distortions that make conventional views quietly wrong.
Gartner's market definition sets four mandatory capabilities. We use them as the procurement test, and we run the first baseline before you commit to anything.
Platform-agnostic. We work with what you already own — Celonis, SAP Signavio, ARIS, Microsoft Power Automate process mining, UiPath, Apromore — and where a digital twin of the organisation is the right end state, we design toward it rather than selling it.
We are a small, senior team. We do the evidence work ourselves, we design with your people rather than at them, and we build the internal capability to run conformance after we leave — because a process that nobody monitors reverts within two quarters.
Australia's labour productivity has grown under 0.3% a year since 2015, against a 60-year average of 1.6%, and multifactor productivity fell 0.5% across 2024-25. Process redesign is not a back-office nicety in that environment. It is most of the available upside.
Engagement commitments. Client-specific results provided as references under NDA.
One or two end-to-end value streams mined from your own event logs. Variants, rework, wait states, cost and conformance — the evidence base everything else stands on.
Start hereAll six dimensions designed together, with decision rights tested against real cases, a transition roadmap, and the management-system changes that make it stick.
EnquireOne value stream, redesigned end to end with the people who run it — fit-to-standard first, decision logic modelled, benefit case costed, conformance defined.
EnquireWhich processes are actually ready for automation or agents, which need redesign first, and which should simply be eliminated. Ranked, costed and sequenced.
EnquireNo. We can run a first baseline using extracts from your source systems, and several platforms offer time-boxed assessment licences. Buying a platform before you know which processes matter is the more expensive mistake.
A restructure changes reporting lines. An operating model redesign changes how work gets done — processes, decision rights, information flow, sourcing and the management system. The evidence is clear that changing structure alone produces confident leaders and unconvinced employees.
Before, at the level of intent and decision rights; alongside, at the level of detail. Redesigning in the abstract produces processes no platform supports. Configuring first produces a system that encodes today's inefficiency for a decade.
It is a starting position, not a rule. Every deviation from standard gets priced — build cost, run cost, upgrade drag and the benefit it protects. Some deviations are genuinely worth it. Most are not, and pricing them is how you tell the difference.
Continuous conformance monitoring against the designed model, plus changes to the KPIs and management system so the new way is the rational way for the people doing the work. Without the second, the first just documents the drift.
Yes, and it usually should. Process baselines are most valuable before design freeze, and conformance monitoring is most valuable immediately after go-live — when programs are typically winding down.
A process intelligence baseline takes three to five weeks and shows you, from your own data, exactly where the operating model is failing the strategy.