A company in the European Union — the kind with ERP, digital channels, data pipelines, and "business-critical systems" knotted together like a bowl of spaghetti someone once called an "IT landscape" — has posted a vacancy. They want a Senior Interim CIO. Temporary. Executive. Someone who walks in, "creates calm, clarity, and trust," establishes a "communicable plan forward," and ensures "the right priorities are made."
They are hiring a person, for a fixed window, whose entire job is to walk into a mess, understand it, write a plan about it, and convince other humans the plan is good.
That is not a job. That is a prompt.
What an interim CIO actually does
Strip the executive veneer and the role is almost touchingly mechanical. The job description spells it out:
- "Quickly get into the business, understand the current situation, and create structure."
- "Handle an environment with many parallel initiatives, dependencies, and stakeholders."
- "Create calm, clarity, and trust."
- "Establish a communicable plan forward."
- "Ensure the right priorities are made."
Now translate that into what physically happens between Monday and Friday. The interim CIO books 30–40 interviews with department heads, enterprise architects, and business sponsors. They sit in a room (or a Teams grid) and ask, "What do you own, what's broken, what's blocking you, and what would you do first?" They collect the answers into a spreadsheet. They spot duplicates, contradictions, and dependencies. They draw a roadmap on a slide. They present the roadmap to a steering committee. The committee argues. The CIO revises. They present again. Trust is declared to exist. A plan is "communicated." Priorities are "ensured."
Every one of those steps is a language task with a document output. The interim CIO is a very expensive summarizer with stakeholder management skills and a nice jacket.
The replacement pipeline
Here is how you replace this role at this kind of company — a large consumer-facing organization with a tangled systems estate and too many parallel initiatives — with an AI system. Now.
Phase 1: Ingestion and discovery. Connect an LLM-based agent to the company's existing artifacts: Confluence, Jira, the architecture repository, the ERP vendor's implementation docs, the Slack archives, the last three years of steering committee minutes. You do not need the CIO to "understand the current situation" by asking 40 people. The current situation is already written down, badly, across 200,000 documents. A retrieval-augmented model reads all of it in an afternoon and produces a dependency map, an initiative inventory, and a heat map of where the contradictions live. The human interim CIO takes three weeks to do this and gets it 70% right because they stop interviewing after they get tired. The model does not get tired.
Phase 2: Stakeholder interviews. You still need humans to talk to humans — for now. But the interviews become structured inputs. Each department head gets a 30-minute call with an AI agent that asks the same five questions, follows up intelligently on vague answers, and logs the transcript directly into the system. The "calm, clarity, and trust" part? That is tone calibration. The agent is prompted to be measured, specific, and to reflect back what it heard. It does not sigh, check its phone, or mentally draft its LinkedIn post about the "transformation journey." It is, ironically, more present than most interim executives.
Phase 3: The plan. This is where the replacement gets almost funny. The JD asks for "a communicable plan forward." What is that, in practice? A roadmap slide. Maybe five slides. A narrative around them. Priorities ranked, dependencies noted, a Gantt-ish visual. An LLM with access to the initiative inventory, the dependency map, and the stakeholder transcripts produces a draft in minutes. It cross-references: "Initiative A from the e-commerce team overlaps 60% with Initiative B from supply chain; the ERP migration in Q3 is a hard dependency for both; recommend sequencing A before B and folding B's scope into A's program." That is the exact sentence a human interim CIO would write after six weeks. The model writes it after six seconds because it actually held all the documents in context simultaneously, which no human does.
Phase 4: Communication and "trust." The steering committee presentation is generated. The narrative is tailored to the audience — more financial framing for the CFO, more delivery risk for the COO, more "digital channel maturity" for the commercial director. The model drafts the talking points, the FAQ, the risk register, and the "what we decided and why" memo. A human delivers it — for now — because boards still want to see a face. But the face is reading from a script the machine wrote.
Where it breaks
Without this, the rest is just a provocation.
The interim CIO role has one function that does not automate cleanly: political absorption. When the CFO wants to kill a project that the commercial director's bonus depends on, someone has to walk into that room, absorb the conflict, and broker a deal that both sides can live with without losing face. The AI can propose the optimal trade-off. It cannot absorb the social cost of imposing it. A model can say "fund A, defer B, and restructure C under the ERP program." It cannot sit across from an executive who built their career on Initiative C and tell them, with the right mixture of empathy and authority, that C is now a workstream inside someone else's program. That conversation is 10% logic and 90% ego management, and the 90% is why companies pay interim CIO rates that would make a consultant blush.
There is also a trust asymmetry. A human interim CIO who has "done this at five different organizations" carries pattern recognition that feels credible to a nervous board, even if the patterns are sometimes wrong. The model carries more patterns — it has effectively seen every retail IT transformation of the last decade — but the board does not yet trust a pattern that came from a GPU instead of a gray-haired Swede. The trust deficit is real, and it is the only reason this vacancy exists in human form.
And the "enterprise architecture background" the JD lists as meritorious? That is the part that automates best. Understanding a complex system landscape at an overarching level is literally what LLMs do better than any architect with a whiteboard. The architect draws boxes and arrows and misses three hidden dependencies. The model reads the integration registry and finds them in seconds. The architecture understanding is not the hard part. The hard part is getting two people who hate each other to agree on which boxes to draw.
The real job description
So here is what the company is actually hiring: a human shock absorber for organizational politics, wrapped in a layer of analysis and planning that a machine can now produce faster and, frankly, more thoroughly. The analysis is the pretext. The absorption is the product.
The interim CIO vacancy is a confession. It says: "We have a mess, we need someone to describe the mess in a way we can agree on, and we need a person to blame if the description turns out wrong." The describing is automated. The blaming still requires a human — for now.
If you are the kind of company that posts this role, ask yourself: do you need a CIO, or do you need a really good prompt and someone willing to present its output with a straight face? Because the plan, the priorities, the roadmap, the dependency map, the stakeholder synthesis — that is done. The only thing left for the human in the blazer is the part where the CFO raises an eyebrow and someone has to not flinch.
And if you cannot teach a machine to not flinch, maybe you deserve the day rate.

