AI for the Event Industry: Trends and Prospects.
8 min 
How neural networks are being used in client event preparation — helping retain client loyalty and win new tenders.
The hype around neural networks is gradually subsiding, and the market is looking at them with growing sobriety. Clients are no longer satisfied with a line in a report claiming “use of artificial intelligence.” They want original results. Agencies, in turn, need tools that genuinely optimise costs and produce a stronger product than their competitors.
The Neural Network Is Not a Replacement — It’s a Tool.
Talk of AI soon replacing every professional, with creative concepts being auto-generated in a ChatGPT window, stopped frightening our industry a long time ago. And rightly so.
An event producer builds an entire world for a client. Selecting speakers, curating a menu, designing the spatial logic, crafting the afterparty atmosphere, defining the event’s overall tone of voice — all of this requires a nuanced understanding of people and context that AI simply doesn’t have yet, and most likely won’t have any time soon.
AI’s genuine strengths are the speed of generation — inaccessible to any human — and its power to overcome the blank page. Any starting point that saves time on routine work is already valuable. Those who know how to work properly with neural networks and integrate them into their processes consistently produce better results than those who don’t.
This is precisely why at iMARUSSIA!, we launched a dedicated division for this space: Superium — a meta-agency working with complex digital products, including metaverses and generative AI branding.
The Neural Networks We Use in Practice.
Our day-to-day toolkit spans several categories.
For text: primarily ChatGPT. We tested YandexGPT but set it aside for now — the results lag behind. For visuals: the paid version of Midjourney and Sber’s free Kandinsky, which is actively developing and regularly gaining new capabilities.
Among the most promising tools for event work, we’d highlight specifically: Taskade for project management; Motion, an intelligent AI calendar assistant; and Tome AI for presentation creation. For websites and logos, Framer and Uizard are interesting. For promo reels and mascots: Synthesia.
On the operational side: transcription tools that convert speech to text, and in pilot testing — Tidv.io, a neural network that handles Zoom meeting minutes and structures outcomes automatically.
Where AI Is Already Working in Practice.
Competitive tender preparation.
We’ve been actively using neural networks in tender submissions for around two years. Primarily for generating visual references — especially relevant when proposing a genuinely original idea that’s never been executed before, leaving no existing inspiration to reference. Among our recently won projects, a number of references were created entirely by AI. Several tenders we’ve won were 70–80% written by AI: texts, concepts, illustrations. That doesn’t mean designers are out of a job. Their role has shifted: they now write precise prompts and ensure brand consistency.
Venue and space visualisation for clients.
Another practical use case: generating a visualisation of the future event. Sketches with object placement — stage, seating, key zones — are loaded into the system. The neural network applies the brand, populates screens with content, positions people, and can build a 3D model on request. The client sees a realistic picture in advance and has the chance to “inhabit” the atmosphere before the day itself.
“The neural network doesn’t create the event for us. It gives us speed at the stages where we used to spend time. The time we save goes into what AI genuinely can’t do — directing the details and working with people.”
Anton Savelyuk, Founder of iMARUSSIA!Event activations and branded merchandise.
One specific case study. For the conference “Space of Security: Protecting the Country’s Digital Sovereignty,” we took an unconventional approach to merch branding: we asked several neural networks to generate images on the theme “What Will the Future of IT in Russia Look Like”, selected the eight strongest results, and added the client’s branded frame — Kod Bezopasnosti (Code of Security). At the event, a live T-shirt printing studio was operating: each attendee chose one of the eight artworks and received a branded T-shirt on the spot.
For the first time in a long time, we saw a genuine queue for merchandise. Instead of the planned 500 units, we printed over 600 — fully exhausting the reserve stock. The combination of trending technology, a culturally resonant theme and personal choice outperformed any standard logo-printed merchandise.
Content for social media.
With the departure of Shutterstock and a number of other international stock platforms, the question of illustration became acute. Midjourney and Kandinsky address this partially — but they do address it. A post is written, the brief goes to the neural network, we receive several visual variants, regenerate with a refined prompt, refine further and publish.
The same applies to creating stickers for Telegram. Neural networks generate themed mascots and compact futuristic imagery with ease — and these work both for post-event attendee engagement and indirect brand promotion, since sticker packs are attributed to the company that created them.
An Honest Look at the Challenges.
More than 4,500 companies operate in Russia’s event industry. By my estimate, neural networks are being used — in some capacity — by a few hundred of them.
Several reasons account for this. First, the entry point is genuinely difficult. Adoption happens by trial and error; every company finds its own approach through experimentation. Quality Russian-language courses are scarce; most current and relevant materials are in English. What does work is live knowledge exchange in professional communities and chat groups.
Second, the technology evolves too rapidly. For some, that’s exciting. For others, it’s alarming. According to Avito Rabota data, 50% of Russians support restrictions on AI development on ethical grounds.
Third, neural networks degrade through interaction with people. A Stanford study recorded that over a three-month period in 2023, ChatGPT-4’s accuracy in certain scenarios dropped from 97% to 2%. The primary risk remains unchanged: confidently stated, plausible-sounding facts and research that simply don’t exist.
A separate issue is intellectual property rights. Neural networks freely incorporate third-party material from the web, and the legal framework here remains undeveloped. Lawsuits by artists whose work was used without consent have yet to establish precedent — but that is temporary. Industry professionals would be well-advised to think about this now, rather than waiting for the first high-profile rulings.
“AI is no longer a future trend. It’s a working tool today. The question isn’t whether to use it. The question is how intelligently.”
Anton Savelyuk, Founder of iMARUSSIA!Conclusion:
This evolution is unavoidable. Neural networks are already embedded in familiar products — Adobe integrated AI into Photoshop seamlessly, and the designer simply enters a task in an additional field. Those who win will be the ones who learn to take unexpected solutions from other industries and apply them to their own practice. Emotion recognition systems already exist — and why not introduce a “happiness KPI” as a genuine metric of event success?
Related blogs
Member of the Event Industry Association
a request.
project