Manchester Tech Festival 2024: AI in the Room

Manchester Tech Festival is one of those events that's harder to describe than it is to experience. It's technically a festival rather than a conference, which means it doesn't have one venue, one agenda, or one audience. Over its two-week run in late 2024 (28 October through 8 November) it occupied Victoria Baths, Friends Meeting House, and various fringe venues and social spaces scattered across the city. The core conference programme runs two days. The AI conference runs one. The rest is workshops, socials, satellite events, and the particular kind of informal knowledge exchange that happens when people in the same industry spend time in the same city for two weeks.

I find it a more useful temperature gauge than the big London events, partly because the audience is different. This isn't the vendor showcase of ExCeL or Olympia, and it isn't the C-suite of NRF. It's practitioners: people who actually build things, lead teams, and manage the daily reality of deploying technology in organisations. When the AI conversation at Manchester Tech Festival shifts from something that happens in a dedicated "AI in business" breakout to being embedded across the software engineering, product, UX, data, and leadership tracks, that's a meaningful signal about where AI actually sits in practitioner thinking.

In 2024, that shift was visible.

The AI Conference: 31 October at Friends Meeting House

The dedicated AI conference ran on Halloween, which is a scheduling choice that says something about the organisers' sense of humour. It was also the first AI conference MTF had ever run. A recognition, really, that the conversation had grown large enough to merit its own day rather than being squeezed into a breakout slot at the main festival.

Friends Meeting House in central Manchester is one of the better venues in the city for a focused technical day: smaller, more intimate than a conference centre, with the atmosphere of a place where people come to think rather than to be sold to. The format kept things human-scale, which suited the subject matter.

The sessions covered territory from the genuinely technical (ML infrastructure, model evaluation, data pipeline architecture) through to the more strategic (AI ethics, governance, workforce transition). The framing running through most of it was the question of how you move from "AI as a project" to "AI as how we work." That's a genuinely hard organisational question, and the fact that practitioners are asking it out loud, rather than still debating whether to take AI seriously, reflects where the community is.

The use case that generated most discussion in the sessions I attended was AI in operational contexts. Not the flashy consumer-facing applications, but the automation of internal processes, the augmentation of human decision-making in complex environments, and the challenge of maintaining quality and trust as AI systems take on more of the routine analytical work. This is very much the practitioner version of the AI conversation, as opposed to the product marketing version.

The Broader Festival: Victoria Baths, 6–7 November

The core two-day festival moved, as it does, to Victoria Baths. If you haven't been: it's a Grade II*-listed Edwardian swimming baths in Chorlton-on-Medlock, opened in 1906, closed by Manchester City Council in 1993, and slowly being restored by a community trust ever since. Using it as a tech festival venue is one of those slightly mad Manchester decisions that somehow works perfectly. The tiling is immaculate. The acoustics are unusual. The vibe is not like any other conference you've attended.

Content ran across three stages, which meant you had to make choices and would inevitably miss things. The ecommerce and retail threads were present but not dominant. This is a general-purpose tech festival, not an ecommerce event, and the conversations about AI in retail sat alongside conversations about AI in healthcare, financial services, public sector, and software engineering.

What I find valuable about that context is that it surfaces cross-domain patterns. The challenges around data quality, organisational trust in AI recommendations, and the human-AI collaboration design question are not specific to ecommerce. Hearing them discussed in other domain contexts is useful for calibrating which of these are fundamental properties of the technology and which are specific to a particular industry context.

The data quality problem, for instance, is universal. Every domain at MTF that was working seriously with AI was talking about the same constraint: the model is only as good as the data, the data is almost never as good as you'd like, and the gap between "interesting demo" and "production-quality deployment" is almost always about data rather than model sophistication. I've made this argument in the ecommerce context repeatedly, and hearing it confirmed by practitioners in other domains is useful.

The Northern Tech Community and AI

Manchester's tech community has a specific quality that makes it an interesting place to track AI adoption. It's large enough, with a cluster of significant tech employers across Spinningfields and NOMA, to have serious technical depth. It's sufficiently independent of the London fintech and enterprise tech ecosystem that it approaches questions without some of the commercial pressures that shape how AI is discussed in Shoreditch or the City. And it has a strong culture of peer learning and community knowledge sharing that predates the current AI wave.

The Manchester tech community in November 2024 was past the "should we pay attention to AI?" phase and clearly into the "how do we actually do this?" phase. Whether it reaches the "we've done this at production scale and have hard-won opinions" phase faster than the equivalent communities elsewhere in the country is a question worth watching.

I think it might. The engineering culture here has a history of building serious things quietly and then being surprised when people outside the city notice.