Scottish Web Folk conference 2026

Track one, Lecture theatre 3

Doing more with less: using AI to scale content quality

Richard SaundersZengenti

With tighter budgets, smaller teams and growing user expectations, university content teams are under more pressure than ever. This session explores practical ways AI can help improve content quality, accessibility and consistency at scale, enabling teams to work smarter, reduce manual effort and deliver better digital experiences with the resources they already have.

Track two, Lecture theatre 1

Human plus machine: why accessibility testing needs both

Kirsty Sharp and Alex MarshallEye-Able

Most university teams already run automated scans, but automated tools only ever catch a fraction of what matters. The barriers that actually stop disabled students accessing content are found through manual testing, often with AI working alongside it.

This session shows where AI now does the work of a specialist tester, where human judgement still cannot be replaced, and how to build a testing model covering automated scans, manual audit and PDF accessibility that works for a team on a real budget.

Track three, Lecture theatre 2

From one-off request to managed service: the life of a dashboard

Carla Soto QuintanaUniversity of Edinburgh

Following the evolution of a dashboard that started as a one-off analytics request and matured into a managed service as its audience, scope and decision-making impact grew. Carla highlights the challenges and turning points along the way.

Track one, Lecture theatre 3

Building for humans and AI: real lessons from an agentic rebuild

Matthew JonesVepple

University websites drive huge traffic and power LLMs, but their heavy focus on facts and figures wins students' minds, not their hearts. Building an emotional connection is getting harder as applicants consume content via social and third-party AI. We see massive untapped potential in indexing interactive experiences for AI precisely to help prospective students build that connection and picture where they'll live and study.

With 40+ universities and millions of sessions on Vepple, we are currently mid-rebuild of our platform for both humans and LLMs. Creating content for LLMs is a vital topic, but this session focuses purely on what we are learning as we build in the era of agentic development: the agile workflows, AI dev tools, failed experiments, and shipping strategies that are actually working.

Track two, Lecture theatre 1

Rebuilding Clearing: getting the requirements right, then building fast

Jennifer DoyleUniversity of Edinburgh

Edinburgh has spent recent months reconfiguring its approach to Clearing: the messaging, the functionality and the role the prospective student web team plays in the process. Jennifer shares how stakeholder and relationship management helped generate the right requirements, and how AI helped speed up development to hit the deadline. The result is rapidly built but robust and flexible functionality, ready for whatever Clearing throws at them in 2026.

Track three, Lecture theatre 2

How to architect your data for the age of AI

Rob Ryder-RichardsonUniversity of Dundee

Structured content, schema, full data sets and URL structure, and how Dundee tests all of it against their own AI chatbot.

Track one, Lecture theatre 3

Findability: how to get your content found by humans and AIs

Paul Costello and Rob AllenNumiko

Search behaviour is undergoing the biggest shift since the launch of the iPhone. Increasingly students and applicants are asking AI tools, or reading Google's AI Overviews, and are handed an answer before they reach your website. That's the challenge facing all universities right now.

In this session Numiko will share how they've approached Findability for the University of Oxford. Paul and Rob will map out the pragmatic and successful approach to findability Numiko delivered after launching the University of Oxford's new website earlier this year, and show that optimising for AI isn't just a technical bolt-on, it's about good content strategy, clarity and solid tech foundations.

Track two, Lecture theatre 1

The machine builds, the human decides: execution is cheap, thinking is expensive

Mark HigginsonCues.ai

AI-assisted prototyping genuinely works. A small team can now build, in days, things that used to need a procurement process. Here's what we've made and why the hard part moved from what you can build to what you can prove.

Track three, Lecture theatre 2

Testing after you deploy, not just before

Rob McCrearyUniversity of Dundee

As security pressure grows and AI makes it easier to find and exploit software vulnerabilities, web teams need to ship fixes quickly without losing confidence in what reaches production. The challenge is not only proving that code is ready to deploy, but confirming the live environment is still healthy afterwards, where caches, configuration, integrations, routes, permissions and real content all affect the outcome.

Using Dundee's recent post-deploy Behat work as the example, this talk shows how fast advisory smoke tests give confidence straight after a deployment, while scheduled regression tests provide deeper resilience without turning every release into a long blocking process.

Track four, workshop, Room 2F13

Drupal in higher education: a lean coffee

Aaron McHaleUniversity of Edinburgh

Runs 60 minutes, 14:15 to 15:15. If you come to this you will miss the 14:45 talks.

Many institutions run Drupal, so this is a chance for that community to come together and talk about shared challenges. Bring a topic you would like to discuss. The session follows the lean coffee format, where participants vote on which topics to cover.

Track one, Lecture theatre 3

How the University of Dundee made personalisation a team sport

Acquia and VWO AB Tasty by Wingify

Personalisation used to be a one-person job at the University of Dundee: one UX manager built every test, and the rest of the content team waited in line. This session tells the story of what changed.

We share real experiments from the university's website, run on VWO AB Tasty by Wingify, including a small search bar animation that lifted search usage by 80 per cent and a simple microcopy test that surprised the team. We also cover how AI-assisted segmentation let the team target international prospective students in minutes rather than hours, and how, working with Acquia, winning tests get built permanently into the site.

No product demo, just real results, a few surprises, and some live challenges for the room.

Track two, Lecture theatre 1

AI search summaries: a new lens on old content problems

Katie Spearman, Mel Batcharj and Nick DanielsUser Experience Service, University of Edinburgh

Search behaviour has changed. People increasingly get their answers from AI summaries, and website traffic has fallen accordingly. At Edinburgh this has been a chance to look at content design best practice through a new lens, and the team found their evergreen advice still stands. This session covers what changing search behaviour means for how you write, how they trained colleagues to respond to it, and what they learned along the way.

Track three, Lecture theatre 2

What every good stakeholder workshop has in common

Neil AllisonUniversity of Edinburgh

Workshops and engagement exercises take many forms, but a few fundamentals underpin every successful stakeholder session. Neil runs through the essentials that need to be present if a workshop is going to bring people together, solve a problem and deliver something tangible everyone can get behind, with examples of those principles in practice.

Track four, workshop, Room 2F13

Workshop continues

Drupal lean coffee, running until 15:15

Track one, Lecture theatre 3

Customer case study

Terminalfour

Track two, Lecture theatre 1

Creating a content design playbook through a repeatable workshop format

Lauren TormeyUniversity of Edinburgh

A content design playbook helps align your team on how to approach different content challenges. If your team is constantly busy, though, it is the sort of thing that feels like a nice-to-have and never gets done. Hear how one content designer built her team's playbook through a series of workshops in the middle of their busiest and biggest project, using a simple format you can adapt, plus a low-effort way to turn workshop outputs into real guidance.

Track four, workshop, Room 2F13

Moving with confidence: how to build AI capability that lasts

manifesto

Runs 60 minutes, 15:45 to 16:45. If you come to this you will miss the 16:15 talks.

The universities best placed to harness AI will be the ones building capability now, learning from the right experiments and laying the right foundations. In this fast-paced workshop you will work through the ingredients of AI success, then identify a persistent problem at your own institution. What you discover when you try to apply AI to it, the data you would need, where human judgement is essential, the foundations that need strengthening, is exactly what an effective AI practice is built on. Come ready to collaborate, whether you are just getting started, feel stuck, or are ready to scale.

Track one, Lecture theatre 3

Accessibility: moving from measurement to remediation using AI and people

Gavin ColborneLittle Forest

For many universities, measuring accessibility compliance has become routine. Acting on it is the hard part. Manual remediation cannot keep pace with the volume of content created every day, which leaves institutions exposed to legal risk and students facing barriers that never get fixed.

This session sets out a framework for closing the gap between finding a problem and fixing it, by pairing AI with human expertise. AI brings speed and pattern recognition, people bring judgement on the cases that need it, and together they turn an overwhelming audit into a manageable pipeline. You will leave with:

• a blueprint for moving from continuous auditing to human-in-the-loop remediation
• a clear sense of where AI does the work well, and where human judgement is non-negotiable
• practical ways to prioritise high-impact content and allocate people alongside AI tools

Track two, Lecture theatre 1

Digital Unscripted

Stratos Filalithis, University of Edinburgh, and Kris Purdy, University of Glasgow

Pull up a chair. No slides, no case studies, no rehearsed takeaways. Just two digital leaders from two of Scotland's leading universities having an honest conversation in front of an audience.

Kris and Stratos have something in common: a web estate that grew for decades while nobody was quite looking. Thousands of websites, hundreds of domains, ungoverned content. Both are now picking through what they have inherited and deciding what comes next. There is no tidy conclusion waiting at the end of this one, but if you are navigating something similar, it might be exactly the conversation you needed to hear.

Track three, Lecture theatre 2

Course journey reimagined

Paul LeighUniversity of Limerick

Limerick's course pages had grown organically inside the university's structures, leaving prospective students with an inconsistent, broken journey. Undergraduate and postgraduate offerings looked different and sat on two different content types. The redesign was the chance to fix it. Driven by UX research, the team:

• built new course pages around what users actually need
• merged two content types into one
• brought all recruitment information into ul.ie/study
• built a course search with faceted filters and areas of interest
• added side-by-side course comparison

Track four, workshop, Room 2F13

Workshop continues

Moving with confidence, running until 16:45