Interim and fractional CDAO in short
An interim or fractional CDAO gives data and AI one executive owner: data governance, the platform the data runs on, and the AI use cases the business takes into production. The interim decides which data is trusted, who owns it and which uses are worth running, until a permanent CDAO holds the seat.
Give data and AI one executive direction.
AI initiatives depend on data that has its own owners, platforms and constraints. The interim CDAO connects those foundations to the business uses the organisation chooses to pursue.
The role leads data governance, platform priorities and the oversight of AI evaluation and use. Investment choices address both the intended application and the conditions needed to operate it.
When to bring in an interim CDAO
- Data sits in silos: each function keeps its own figures, definitions differ from one report to the next, and nobody has the authority to settle which version the business runs on.
- The business has an AI ambition but no platform to carry it: use cases are piloted one at a time, on data extracted by hand, and none reaches production.
- The technology function has an owner, but data and AI decisions are spread across it, the business lines and a few specialists, with no executive answerable for them as a whole.
- The data leader has left, and governance, platform priorities and the oversight of AI use are waiting for someone with the authority to decide them.
When it is the wrong answer
- The direction for AI is decided and owned internally, and what is missing is the capability to build and run a model or an agent. Scale with AI brings that capability without taking the seat.
- The data direction is owned internally, and what is missing is the capability to set the governance, build the platform and the pipelines or give decisions trusted indicators. Data brings that capability without taking the seat.
- Nobody yet agrees on what is wrong: the data, the platform, the organisation or the ambition itself. Start with an Assessment, then decide whether a seat is needed.
- Data is one part of a wider technology function with no owner, where architecture and delivery weigh as much as data. The CTO seat is the better frame.
What the CDAO holds
- Data governance: who owns each domain, which definitions the business reports on, and the quality a dataset must reach before a decision rests on it.
- The data platform: its priorities, what is built, bought or retired, and the order in which sources are brought onto it.
- AI use cases: which ones go to production, which stop at the pilot, and the conditions each must meet to keep running.
- The oversight of AI evaluation and use: how a model is tested before release, who watches it once it runs, and when it is withdrawn.
- Investment choices that fund both the intended application and the data, platform and people needed to operate it.
The first 90 days
Days 1 to 30
- Settle what the mandate fixed at entry: the data and AI decisions in scope, the results that will count as success, the reporting line and the conditions for handing the seat over.
- Map the data as it is used: sources, owners, conflicting definitions, and the AI pilots already running with the data they depend on.
Days 31 to 60
- Name an owner for each critical data domain, and agree the definitions the leadership team will report on.
- Sort the AI use cases into those that go to production, those that wait for data and those that stop.
Days 61 to 90
- Set the platform priorities and the evaluation rules every model passes before and after release.
- Hand domain ownership and use case decisions to the people who will hold them after the interim.
What you keep
- A data governance model: domain owners, shared definitions and the quality rules decisions rely on.
- A platform roadmap with its priorities and the sources it brings on, in order.
- A portfolio of AI use cases, each with a decision to run, wait or stop, and the reason for it.
- Evaluation and monitoring rules for every model in use, each with an owner.
- A data and AI organisation whose leads decide without routing through the interim.
We leave when the function can run without us: an identified permanent owner, a roadmap they hold, and a team that no longer routes through the interim.
Questions
- What does an interim CDAO do?
- An interim CDAO takes the executive decisions of data and AI while the seat has no permanent holder: who owns which data, what the platform builds next, which AI use cases reach production and how they are evaluated. The interim answers for those choices in the client's reporting line and leaves a function that decides without them.
- When is an interim CDAO the wrong answer?
- When the direction for AI is decided and held internally and only the capability to build is missing: Scale with AI fits better. When nobody agrees on the problem, an Assessment comes first. And when data is one part of a technology function with no owner, the CTO seat is the better frame.
- What does the CDAO hand over at the end?
- A governance model with named domain owners and shared definitions, a platform roadmap, a portfolio of AI use cases with a decision on each, and evaluation rules for every model in use. The permanent CDAO inherits a function whose leads already take those decisions, not a list of pilots waiting for an owner.
- Why one executive for data and AI?
- AI initiatives depend on data with its own owners, platforms and constraints. Split between two leaders, a use case is chosen on one side while the data it needs is prioritised on the other. One executive direction connects those foundations to the business uses the organisation chooses, and answers for both.
- Does the CDAO build the AI applications?
- No. The CDAO leads data governance, platform priorities and the oversight of AI evaluation and use, and decides which applications are worth building. Engineering teams or partners build them. When the organisation needs that implementation capability itself, with its direction already set, Scale with AI in the Expertises brings it.