To a new UX researcher, I’d say: ditch the storytelling. Focus on building domain expertise.
In an AI-first world, the core activity of value generation is shifting from presentation to precision and implementation.
AI Key takeaways:
As enterprise AI moves in-house, traditional UX storytelling and performative research reports become obsolete.
Research deliverables will split into a strict binary: clean, machine-ready data feeds or fully detailed, implementation-ready solutions.
The external researcher’s value shifts from synthesizing information to extracting verified, real-world data that prevents the internal model from becoming an echo chamber.
Domain expertise in rigorous data extraction and the ability to implement innovative, realistic solutions will replace narrative packaging as the ultimate credibility signal for clients.
Index
The local model era is coming
Research will not be talking to human anymore
The binary output
Analysis moves in-house
Where the value actually is
What happens when you ignore this
The future of UX researchers, UX designers, and clients officers
Storytelling will fade away.
Expertise will make or break consulting.
In an AI-first world, the consulting industry is not disappearing, but the core activity of value generation is shifting from presentation to precision and implementation.
Why? As enterprise AI models move in-house and master internal company context, they will take over strategic analysis and render the traditional, narrative-driven consulting report obsolete.
My forecast? Future consulting and ux research outputs will split into two extremes with no middle ground: either clean, structured data designed for machine ingestion, or highly detailed, ready-to-deploy implementation plans.
To survive this shift, consultants must abandon polished packaging and pivot their focus toward deep domain expertise, human-centered data collection, and actionable implementation.
The local model era is coming
Enterprise AI is moving local. Privacy requirements, regulatory compliance, data security, and the need for operational autonomy are pushing industries toward models that run on their own infrastructure. When your model lives in house, your data never leaves. Your IP stays yours and your compliance is built in. Google’s Gemma, open weight small language models, and the rise of edge AI are making this not only possible but inevitable.
“AI sets up Kodak moment for global consultants.”
— Aimee Donnellan, Reuters Breakingviews, October 2025 — Reuters
Once companies operate their own local models, the way they consume external knowledge changes fundamentally. That is where consulting breaks.
Research will not be talking to human anymore
In a model driven world, you won’t be talking to humans. You’ll be talking to in-house LLMs.
Executives already skim to the summary and skip to the recommendations. The research deck was always more performative than functional. Local enterprise models remove the pretense entirely. Data becomes a structured feed optimized for machine ingestion, not human narrative. The report, as a format, loses its reason to exist.
“We’re not going to keep paying $500K for a report that we suspect was generated largely by a machine.”
— Anonymous CIO, Future of Consulting, January 2026 — Future of Consulting
The industry is already building the infrastructure for this shift. What analysts now call Agentic Document Extraction is turning static PDFs into structured, machine readable intelligence. Companies like Box and LandingAI are recovering what the industry calls “dark data,” the overwhelming majority of enterprise information that sits unstructured, unused, and expensive to store. The emerging consensus: if data is not live, it is liability.
And critically, the output now serves the entire company, not one team’s political framing. A model fed with clean enterprise level data does not, cannot play politics. It does not omit findings to protect a department’s narrative (obviously, this will create some frictions with personal agendas at the beginning). When prompted for information internally, the prompter will do the curation.
Storytelling was a fig leaf.
When the report goes away, so does the need for storytelling.
Aesthetic curation and emotional narrative are dead, but data structuring and information architecture are the new gold standards. Politics shifts from the presentation layer (the slide deck) to the data governance layer (what gets fed into the model) and the prompt layer.
The product of a consulting engagement should be a living solution that the client can use, not a document about the solution.
The binary output
My guess? Research splits into two outputs with nothing in between.
Either you deliver clean, raw, unanalyzed data structured so the client’s model can contextualize it against everything else the company knows.
Or you deliver fully detailed, implementation ready output precise enough to be deployed without further interpretation. A table with copy/paste-ready content in the last column. (Almost) final-stage product design prototypes. A roadmap with specific steps, owners, and timelines.
Both ends of this split are expert work.
The middle ground of high level trends accompanied by vague recommendations serves neither the model nor the human. It is redundant. The traditional deliverables that consulting was built on are increasingly insufficient.
And speed is the reason this middle ground will not survive. AI does not mean less work in less time. It means more work because each unit takes less time. When research gets faster, you do more of it. You do not spend the saved time writing a prettier report. The report was a bottleneck disguised as a deliverable, and the acceleration of research cycles will expose it as one.
Ten focused data collections are worth more than one polished deck with high level trends and a couple of illustrative quotes. Especially if the data is to be read by a machine.
Analysis moves in-house
Here is the uncomfortable truth; a client’s enterprise model will always outperform an outside consultant at strategic analysis. Not because the model is smarter, but because it has more context. It knows the company’s data, history, internal dynamics, and goals.
More critically, and it needs to be crystal-clear for consultants: pure analytical strategy naturally migrates in house to the client’s model.
What stays on the consulting side is either expert data collection or expert implementation. Everything in between gets absorbed.
This paradigm is redefining the external consultant as the API to the real world.
Where the value actually is
The researcher’s job does not shrink, far from it. The analytical layer gets decoupled, and the actionable knowledge layer becomes the primary value. So does an advanced degree in behavioral science, sociology, philosophy, or economics. Basically anyone whose training sharpened their ability to understand how people think, decide, and behave, and what the mechanisms of the solutions will be.
In this new paradigm, domain expertise rises sharply in value. A medical specialist asks sharper clinical questions. A marketing expert catches nuances others miss when talking about a campaign. A designer can make better usability solutions on a product. A veteran researcher with hundreds of IDIs under their belt is better at reading a room during an interview.
You need domain knowledge to understand what the client actually needs versus what they are asking for. Building rapport with study participants. Asking the right questions. But also, animating workshops, leading group ideations, and helping build consensus across departments.
The quality of what you gather, not how you package it, becomes the differentiator.
But expertise here is not only technical or domain-specific. I said it already, and others did too, in an AI-first world, culture, humanities, critical thinking rise too, across domains. The ability to understand context, read people, ask the question nobody else thought to ask. That is expertise.
What happens when you ignore this
What happens in an AI world when storytelling is the primary driver over data collection expertise?
Deloitte provided the case study twice in the same year. First, an Australian government welfare report was found to contain fabricated AI generated citations, nonexistent footnotes, and a fabricated court quote. Then a Canadian healthcare report surfaced with similar issues.
“The use of standardized, cookie-cutter reports by consultants that are shopped to different municipalities with no regard for local context or history is yet another example that there often isn’t much substantive research or analytical rigour behind the curtain.”
— Canadian Centre for Policy Alternatives, December 2025 — CCPA
AI did not create the problem. For Deloitte, the packaging was always doing more work than the analysis. When the packaging got automated, the absence of expertise underneath became visible.
The future of UX researchers, UX designers, and clients officers
The expert gathers the data and produces the output. They are the credibility signals that the client is looking at when deciding if your input is valuable. The leadership seat belongs to them.
Keeping up won’t be as easy. In such dynamic and changing landscape, domain expertise is hard to maintain, demands constant effort, and personal investment.
The previous model did not demand deep domain expertise from everyone in the chain, and that was fine for its time. The new model does. That is not a verdict on anyone but a shift in what the market rewards.
Relationship builders and facilitators remain essential, but in a supporting capacity that enables experts to deliver, not as a filter between the expert and the client.
For a lot of us who built management skills, it’s not nice to hear, and we’ll have to let go of the ego.
For professionals who built their careers on narrative and packaging, two paths remain:
The first is to become an expert in something, an expert that ideally can prescribe actionable solutions. That does not have to mean a niche domain or technical expert (medical, legal, design, etc). It can mean research itself. The craft of extracting clean, trustworthy, model-ready data. Knowing which questions to ask. Reading a room. Understanding what to probe and what to discard. That is a real expertise, and it leads. The critical thinking does not disappear but relocates instead. From how you present the data, to how you gather it right in the first place, and what to do with it.
The second is to lean fully into the human side. Client relationships, facilitation, conversation and empathy. That is valuable and it has a place. But it is a supporting role.
Realistically, it won’t be immediate, internal politics and personal career trajectories will be in the way, at first. But it will happen. The center of gravity will shift. From storytelling to expertise. From reports to structured data. From human translation to machine ready precision.
The Child and Death, Edvard Munch1899