What should an AI transformation consultant actually do?
A real AI transformation consultant redesigns how your business works, then deploys technology into the redesigned work — not the reverse. AI is a business transformation, not a technology rollout, and the order matters: redesign the operating model and the workflows first, then bring the technology to them.
The technology is the smaller part of the job. In a successful enterprise AI program the majority of the effort — roughly 70% — is people, process, and organizational change: redesigned workflows, new roles, governance, training, and adoption. A partner who only stands up models and tools has solved the easy part and left the 70% that actually determines whether the investment pays off.
Boutique firm, large consultancy, or AI tool vendor — which do you need?
For enterprise transformation you want senior practitioners who are model-agnostic and accountable for outcomes — which points toward a boutique over a leveraged large-firm team or a tool vendor. The three options optimize for different things, and the differences are practical.
A tool or platform vendor sells you their stack; their incentive is to fit your problem to their product. A large consultancy brings scale and brand, but the people who win the pitch are rarely the people who do the work — engagements are delivered by leveraged, often junior, teams at a markup. A boutique keeps senior practitioners in the room, carries less overhead, and owns the result directly. The trade-off is raw scale, which matters less than seniority for the strategy, workflow-redesign, and change work where transformations succeed or fail.
What questions should you ask before hiring one?
Ask questions that separate a transformation partner from a tool implementer. The most revealing ones:
Do you start with our business and workflows, or with a tool? Are you model- and platform-agnostic, or tied to a specific vendor? Who actually does the work — the senior people in this room, or a leveraged team we have not met? What is your methodology, end to end, from strategy through sustained operation? How do you handle the people and change side, which is most of the work? How and when do you measure success? And can we see your methodology before we engage?
The best answers are specific and business-first. Vague, tool-centric, or "trust the process" answers are themselves a signal.
How do you evaluate their methodology?
Look for a defined, sequenced methodology that runs from strategic thesis through deployment and continuous operation — not a bag of disconnected use-case pilots. A credible partner can show you the whole arc: how strategy becomes resourced imperatives, how work gets redesigned around AI, how it is deployed and governed, and how the organization keeps improving after go-live.
The strongest signal is transparency. A firm that publishes its methodology in the open lets you judge the thinking before you spend a dollar. If you cannot see how a partner works until after you have signed, that is worth noticing.
What are the red flags?
The reliable warning signs all share one root: technology-first thinking. Watch for a partner who leads with a tool or platform rather than your business; who is tied to a single model or vendor and calls it a recommendation; who frames transformation primarily as headcount reduction rather than capability building; who accumulates pilots with no path to scale; who staffs the work with junior teams after a senior pitch; or who has no defined methodology and no plan for the people-and-process majority of the effort.
Any one of these predicts the same outcome — an expensive proof of concept that never becomes a business result.
How Plaster Group approaches it
Plaster Group is a boutique, AI-native solutions consulting firm — owned and operated in Seattle since 2008 — that has served enterprise clients across more than 13 industries for nearly two decades. We are model-agnostic and platform-independent: our recommendations follow the work, not vendor allegiance.
We publish our complete five-level AI Business Transformation Methodology in the open, so you can judge the thinking before you engage. Start with the methodology itself, see how we deliver, or reach out when you are ready to talk.
the five-level methodology, how we deliver, the 27-article series, or start a conversation.
Frequently Asked Questions
What does AI business transformation cost?
It varies with scope, but the honest answer is that cost tracks the size of the business change, not the price of the technology. Most of the budget should go to the people-and-process work — workflow redesign, new roles, governance, and adoption — rather than to models and tools. Be wary of fixed-price "AI pilots" priced like a software project; they usually scope out the 70% that makes transformation actually work.
How long does an AI transformation take?
There is no single go-live. Enterprise transformation runs in waves rather than as one project. Expect the first wave to take longer, because the organization is building new capability while doing the work; subsequent waves accelerate as playbooks and patterns form. Treat any promise of a fast, one-and-done rollout with skepticism — the technology is quick to stand up, but the organizational change around it is the real timeline.
What is the difference between a CAIO and a CSO?
They are complementary, not interchangeable. The Chief Strategy Officer (CSO) sets where the business is going and frames the strategic thesis. The Chief AI Officer (CAIO) owns how AI transforms the organization to get there — building and running the transformation capability, resourcing the imperatives, and ensuring the people-and-process work actually happens. The CSO defines the destination; the CAIO builds the vehicle.
Should we build an in-house AI team or hire a consultant?
Both, in sequence. A good consultant accelerates the first waves and transfers the methodology, with the explicit goal of leaving you with an in-house capability rather than a dependency. Be cautious of any partner whose business model requires you to keep paying them to operate what should become your own function — the aim is to make yourself independent, not to install a permanent outsourcer.
Do we need to fix our data before starting AI transformation?
Not entirely, but data readiness is a real constraint you cannot ignore. You do not need perfect data to begin, but you do need a clear-eyed assessment of what is siloed, missing, or ungoverned, and a plan to address it as part of the program. The failure mode is treating data as either a non-issue or an all-or-nothing prerequisite; the right approach is to remediate it in parallel with the transformation, prioritized by the workflows you are redesigning.