2024 AI Predictions on UX Research Review

I’m not an expert, but I do know AI predictions are notoriously conservative… by ME

In February 2024, I wrote an article "Vesperance for UX Researchers" outlining possible scenarios for UX researchers, the effect of AI on digital products, and our lives. One year on, let's review those predictions and see what has transpired since.

I'll note, most predictions weren't given a year, in fact most claims had no date at all, yet some have come to fruition. Some 2025 predictions?have already been achieved, while others are on the path to be achieved.

I'll label them as such:

• ? Achieved
• ?? On-Target
• ?? Delayed
• ?? No Soup for You!

Date Not Predicted, Achieved in 2024

Without a date, here are 3 prediction achieved in 2024.

? Ongoing Tension between UX and Product Teams

• Prediction: The already-strained dynamic between Product and UX could accelerate AI’s adoption for research, which might further devalue traditional UX researcher roles.
• Comment: Wasn't meant to be a prediction so much as commentary as it was already happening and overly obvious. However, the tension is starting to be more widely voiced and thus acknowledged.
• Supporting references:

1. The UX Reckoning: Prepare for 2025 and Beyond (NNgroup – January 2025)

• Outcome: This is actually more "ongoing" than "achieved". UX practitioners are advised to shift their professional focus and ethos – essentially our integrity – conceding, we lost.

? Use of AI-Generated Participants

• Prediction: AI-generated “synthetic users” that mimic real individuals could replace or supplement real participants.
• Comment: Acceptance is growing in supplementing humans, but with caution in terms of replacing them.
• Supporting References:

1. Synthetic Users is currently leading this space, citing 22 validation sources, including the study “Generative Agent Simulations of 1,000 People” ( ARXIV – 2024), which claims an 85% accuracy rate in replicating real survey responses.
2. Synthetic Users: Revolutionising UX Testing and Digital Performance (Entropik – November 2024) recommends combining synthetic and real user testing.
3. Synthetic Users: If, When, and How to Use AI-Generated ‘Research’ (NNgroup – June 2024) endorses limited, responsible use.
4. Using ChatGPT in Tree Testing: Experimental Results (MeasuringU – April 2024) warns that ChatGPT is non-deterministic but can be used meaningfully in research.
5. AI-driven Personas to Lead Market Research Data by 2025 (CFOtech – October 2024) found that 71% of surveyed researchers expect synthetic responses to soon dominate data collection.

• Outcome: As predicted, the industry is justifying humans being supplemented with synthetic users.

? Achieving AGI

• Prediction: AGI (Artificial General Intelligence) would emerge sooner than expected, possibly within the year.
• Comment: Whether your goalposts are on wheels or firmly in the ground will determine this outcome. If it is on wheels, then AGI may never be achieved. As for me, AGI was achieved in September 2024 with OpenAI's Strawberry, o1 Preview model. If not o1, o3 in December 2024 should have kicked a goal.
• Outcome: Although there's no universal agreement on the definition for AGI, if o1 was available pre-GPT, I reckon they would say o1 is AGI. So, I'm claiming it, as "Achieved".

2024 Predictions

My lone prediction for 2024 didn't happen... thankfully!

?? Lay-Offs and Industry Restructure

• Prediction: Ongoing and future lay-offs in tech suggest a changing landscape where UX research is increasingly seen as optional in the coming months.
• Comments: Well I am happy to have gotten this one wrong! Unfortunately, there was no increase in UX research opportunities either. UX has always been seen as a luxury, blamed for causing delays, and first on the list to be cut. The digital industry is happy to move on and is actively doing so.
• Outcome: "No Soup for You!" but I didn't want to eat that soup. Will 2025 be any different? You'll find out in my 2025 predictions.

2025 Predictions

These make up my previous 2025, and on a probable trajectory, predictions:?

?? AI Agent?Swarms?

• Prediction: UX researchers will shift to managing multiple AI agents to accomplish UX-related tasks, effectively relegating the researcher to a “swarm managers.”?
• Comment: This hasn’t progressed as quickly as I envisioned but it’s still early days as it is only just 2025, that’s with respect to UX research. Outside of UX, AI agents have been in use at least since 2023. Now, AI agents are the low hanging fruit and everyone is “predicting” 2025 will be ‘the year of AI agents.’?
• Outcome: AI agents were in common use in 2024, but largely only within the highly technical circles. Hints of their use within UX research?—?other than from me?—?indicates they are on the way and thus On-Target.?

?? Mixture of Experts (MoE) Technique

• Prediction: Multiple specialised AI models or “experts” will collaborate to accomplish tasks, likely gaining traction in UX research.
• Comment: I mentioned MoE with AI agents in 2025 and it will be difficult to find anyone who thinks otherwise at this point.
• Outcome: This one is definitely On-Target!

?? Short Shelf-Life for UX Researchers?

• Prediction: Up to 85% of current UX researchers may face redundancy within 12–18 months due to AI’s rapid evolution.?
• Comment: August 2025 will be 18 months. Still 6 months remaining, and I still believe this will happen. However it may take an additional 4–6 months to come to fruition, making it delayed until 2026, but it is on target.?
• Outcome: Too early for certainty but I’m going to mark it as On Target.?

?? Remaining 15% in Slower Sectors

• Prediction: A smaller portion of researchers may continue in slower-moving industries (e.g. government, enterprise, education, health, or corporates doomed to fail).
• Comment: This cannot be determined until we see the herd transition in my previous prediction.
• Outcome: My prediction remains, but too early for certainty and thus On Target.

?? Temporary ‘UX-AI Manager’?Roles?

• Prediction: Some UX researchers could transition into roles managing AI in UX contexts (e.g. UX-AI Managers, Swarm Managers), though these positions may only bridge the gap before they too become unnecessary.?
• Comments: We are starting to see professionals change their titles to “AI-UX Researcher” or similar. There are articles discussing the need for UX research roles needing to evolve to incorporate the management of AI agents.?
• Supporting References:?

1. UX Roundup: AI Research Window | AI Skill Substitution | World Economic Forum Predicts More UX Jobs | AI-UX Consultant | Bad Bird Lip-Synch (Jakob Nielsen?—?January 2025)?
2. Ioana Teleanu’s LinkedIn post (January 2025)?

• Outcome: This one might be considered achieved, but I’ll market it as On-Target.?

?? Role Displacement?

• Prediction: Traditional UX research tasks could be replaced or heavily augmented by AI.?
• Comments: This is starting to happen and we are already advised to shift our career space.?
• Supporting references:?

1. The UX Reckoning: Prepare for 2025 and Beyond (NNgroup, January 2025)?

• Outcome: Unfolding in real-time, On-Target…?

Date Not Predicted, Progressing

Here are several predictions which I still stand behind as likely scenarios to happen but I didn't say when.

?? AI-Driven Real-Time Monitoring and Updates

• Prediction: If I'm being honest, this is really a two prong prediction around what AI will provide in real-time: Monitoring, quickly fix issues, prevent security issues, and the like;Adjust designs, repairing UI issues, database dramas, and ultimately reducing the role of human researchers in ongoing product iterations.
• Comment: Unknowingly, the 1st one already existed and coined as "AIOps Tools" (Product examples below). These platforms employ advanced analytics and machine learning to monitor and manage applications and hardware in real-time. The 2nd one is in its infancy and due to be a huge new product offering.
• Products available (just a few):

1. Freshservice https://www.freshworks.com/
2. ManageEngine https://www.manageengine.com/
3. Datadog https://www.datadoghq.com/
4. Hexometer https://hexometer.com/
5. Dynatrace https://www.dynatrace.com
6. LogicMonitor https://www.logicmonitor.com/

• Outcome: Although no prediction date was given, and one already exists, the 2nd is On-Target to be a thing but more likely in wide-spread use in 2026.

?? Meta, Google, and Syntheticusers Products

• Prediction: These platforms are working on or have already introduced synthetic user technology. This could lead to a future where research with real users is reduced or overshadowed by AI-based proxies.
• Platforms:

1. Meta AI Studio Chatbot - is in publicly available Alpha, focused on businesses and public personalities. Nothing new about Facebook profile data being used.
2. Google Ellmann - no new news found since December 2023.
3. Synthetic Users - actively used product with the intent of supplementing real users.

• Supporting references:

1. Meta AI Studio Chatbots - https://ai.meta.com/ai-studio/
2. Google AI's Ellmann - no public URL.
3. Synthetic Users - https://www.syntheticusers.com/

• Outcome: 2 of 3 products are known to be in progress, so it's On-Target.

?? No-Code Tools (e.g. Make Real)

• Prediction: The ability to create and iterate functioning digital products in minutes will lessen the perceived need for upfront UX research.
• Comments: The number of tools in this space has exploded since the release of TL;Draw's Make Real (https://makereal.tldraw.com/) and we can expect them to continue to grow. This is an unstoppable future for products as design and development is democratised.
• However, these products are much more powerful and go much further than simple text-to-product (TtP) platforms. Also commonly referred to as autonomous coding agents, you can now create & deploy websites & apps to a URL, connect to your data-store, scale, monitor, handle automations, build internal tools, data pipelines and do so in any programming language. No setup, downloading, or coding knowledge required. Just upload a picture/drawing or tell it what you want, iterate through chat, perform usability testing, and launch.
• What used to take weeks, months, or years and cost $1,000s to $100,000s is now a 'done for you' system from seconds to hours depending on the product. Some tools are free to use while others have monthly subscriptions less than $20 per month.
• Products available (just a few):

1. Bolt.new https://bolt.new/
2. Databutton https://databutton.com/
3. Lovable https://lovable.dev/
4. Replit https://replit.com/
5. v0 https://v0.dev/

• Outcome: I cannot say this has happened but it will no doubt be the future and thus another factor in the death of UX research. The fact these products exist at all is well and truly driving innovation and democratising both design & development. On-Target!

?? ROI Arguments Becoming Moot

• Prediction: Common ROI justifications for UX (e.g. early discovery of issues reducing costs and support tickets, preventing brand damage, etc.) may lose impact as AI-driven solutions take over those tasks.
• Comments: Even without AI this was starting to be the case due to the perpetuation of the Product Team movement.
• Outcome: This will continue until GUI products are not pre-designed. I'm tagging this one as On-Target.

?? Decreasing Emphasis on Traditional Usability

• Prediction: Younger users adapt to half-baked products more readily, and AI will be able to compensate for flaws in real time.
• Comments: Research continues to show younger users are more interested in visual design and personalisation, while older users still value ease of use and the need for accessibility. However, as the population continues to shift from those born without tech to those never living without it, the shift is accelerating.
• Outcome: This prediction will continue to evolve without a fixed point to establish whether the prediction is correct or not. So I'll tag it as "Delayed".

?? Future of Personal AI and Products

• Predictions: The following predictions are being grouped together as they are quite similar as they are dependent on one another.

1. Highly Personal AI Assistants: Each of us will have a dedicated AI that caters to our specific needs, eliminating the market for one-size-fits-all products and services through personal educator, doctor, life coach, etc.
2. Customised ‘Half-Baked’ Products: Products might be intentionally released as bare-bones frameworks, leaving our personal AI to tailor them to our exact perfect needs and preferences.
3. Virtual AI Communities: AIs may collaborate in shared networks to exchange ideas or ‘modules,’ shifting how new digital solutions are conceived and distributed.
4. Possible New Commerce Models: Revenue might come from selling AI modules, ideas, and ‘life experiences’ (e.g. fully immersive or dream-based scenarios) rather than the conventional product marketplace.

• Outcome: No date was given for an outcome. More on this in my upcoming 2025 predictions.

Conclusion?

Whether you think AI progress is good, bad, or ugly, I would say these potential scenarios I outlined are still a possibility.

I’m almost finished with my new 2025 and beyond predictions, which I will publish soon. I’ll add a comment once published.

Until then, feel free to tar & feather me…?