Phil McMinn

Director of Digital Marketing

AI: What’s Next for Nonprofit Digital Marketing?

9 mins read

Over the last six months, the Torchbox team has been laser-focused on looking ahead at the implications that the “arrival” of AI is potentially going to have on the world of digital for nonprofits. The focus of our work and our progress is being documented at by our amazing Innovation team.

In our Digital Marketing team, we have an “early AI adopters” group (made up of staff representing paid media, SEO, email and data) who meet weekly with this innovation team to collectively identify and solve smaller challenges that are centred around real world Digital Marketing challenges our charities face . It’s clear already that there are major shifts for marketers coming.

Below, our three Digital Marketing ‘Heads’ outline the view from their individual teams.

An AI generated image of three women looking into a space vortex as a metaphor for looking into the future

AI helped create an image of our digital marketing leads future gazing. Just for fun.

Emma Bennett, Head of SEO

Search Generative Experience (SGE)

In May 2023, Google announced one of the most transformative developments since its inception. Google’s Search Generative Experience (SGE) integrates AI technology directly into the search engine, allowing users to generate AI “snapshots” and ask conversational follow-up questions.

We expect this change will fundamentally impact the way people search for information, and our focus is now on ensuring that our clients are as visible as possible in this new era. The main areas to be aware of are:

  • AI-Powered Snapshots - These snapshots at the top of the results are AI-generated and serve as what Google describes as a ‘jumping-off point’ from which people can explore web content.
  • Conversational mode - People can “ask a follow-up” question or select a suggested next step - this will launch the new conversational mode.
  • Vertical experiences - There are major opportunities within SGE for charities with e-commerce or local verticals, as these types of results have huge scope for comparing features within the AI snapshot.
  • Advertisements - Google describes Ads as a ‘vital piece of how the web works’. Search ads will continue to appear in dedicated ad slots throughout the page.

For charities to adapt to SGE, here are three things that we think will be more important than ever:

  • Everything we currently know about good SEO will still be relevant for being discoverable in SGE. Google states they have “trained these models to uphold Search’s high bar for quality... they rely on our hallmark systems that we’ve fine-tuned for decades”. These systems are those we’ve been working closely with for many years and our search marketers at Torchbox are super excited to adapt to these new developments.
  • Experience, Expertise, Authoritativeness & Trustworthiness (E-E-A-T). Google has made the high bar clear within all documentation on SGE so far, and demonstrating why your information is authoritative enough to be a source will continue to be a major factor in organic search visibility.
  • Technical criteria including markup, backlinks & optimised page titles. Helping Google to better understand the content on your site through structured data, demonstrating your status as a trusted source through citations and optimised, enticing titles (to offset the reduced or missing descriptions we’ve seen in some testing) will all contribute to visibility.

SGE’s impact on health content

Our clients include major health organisations such as NHS England, MS Society, Stroke Association and Diabetes UK. As such, a major consideration for our SEO team is on how SGE, Bing’s ChatGPT-integrated search engine and other AI tools are returning health information.

Many of our charities have YMYL (“Your Money, Your Life”) web pages, containing information related to wellbeing, finances, health and civic information. Our initial assumption was that YMYL queries would be largely off the table for AI snapshots, given Google’s statement: “We’ve also applied additional guardrails, like limiting the types of queries where these capabilities will appear.”

We’re not seeing Google state that health information is completely off-limits for SGE, rather that it will be “more selective” in using SGE for YMYL queries, heavily testing before rolling out and providing more factual than conversational answers.

However, in early releases in the US recently, answers were appearing for YMYL topics like health, financial and legal queries – with many disclaimers.

We know that high-quality, highly authoritative information is as crucial as always, with Google raising the stakes in order to appear in these AI snapshots: “We hold SGE to an even higher standard when it comes to generating responses about certain queries where information quality is critically important.”

It’s worth noting that SGE will not be used for sensitive, controversial, political or dangerous queries, which is relevant for charities providing information to people in crisis. Google describes these queries as those that “indicate a vulnerable situation – for example, on self-harm queries”, and explains that their systems will show trusted interventions (such as emergency telephone numbers) at the top of the results.


Another major upcoming feature in Google Search is the new Perspectives tab, which brings social media, social video and forum content directly into the search results. Google has stated that when users are “looking for something that might benefit from experiences of others” they’ll see the perspectives filter appear, showing ‘firsthand’ or ‘expert’ information from wider digital channels.

This is something we’re extremely excited about at Torchbox because many of our clients have a wealth of rich video, social and community content across their digital estates. The relationship between search, social, video & communities will become more important than ever before, and leveraging the more tangible and direct link between these channels will be a huge opportunity for many of our charity clients.

YouTube SEO, TikTok SEO and leveraging community forums are all new ways in which digital teams can collaborate to provide unique value to wider audiences. These channels provide additional methods of reaching new audiences, and leveraging user-generated content can take the resourcing pressure off of internal content teams. If you’re a nonprofit with a community; Perspectives is an opportunity to increase your reach.

Amina Iqbal, Head of Paid Media

For our Paid Media team, the arrival of AI brings with it a multitude of opportunities to improve and enhance our PPC and social media work. Here are the issues we’re discussing right now as a team.

AI integration with Google Ads

Already, we’re seeing AI functionality in beta within tools like Google Ads, with clear in-platform support coming to support the creation of ads and ad copy. We also know generative AI is coming to Performance Max, the newest kid on the PPC block right now. Soon, it will be possible to use a conversational mode of interaction with the Google Ads interface to describe your business, and the platform will build your ad campaigns for you based on this information.

Big question marks remain as to how effective these tools will be. Google currently offers tools that purport to speed up the setup and maintenance of tools like Google Ad Grants accounts, but no serious marketer would actually use these tools. Whether or not a conversational AI functionality will change that will be something we’ll be watching carefully.

An image showing AI integration with Google Ads

We're already seeing AI functionality in beta within Google Ads

Search Generative Experience (SGE)

SGE will bring significant developments to the paid advertising and PPC space. While Google has mentioned that ads will continue to appear in dedicated ad slots, it remains to be seen whether sponsored ads will be seamlessly integrated into AI-generated responses or if ads will appear independently of these responses (as they currently do). As Google's primary revenue stream comes from its Search ad business, it is a given that they will ensure they have a presence in SGE’s new SERP.

The really interesting element will be how ads integrate: to complement the highly visual style of SGE, might we see more image-led results, more in line with what we now think of as Display ads? How about integrated YouTube ads in the SERPs? One interesting element of SGE is its conversational mode, which allows you to ask follow up questions. At the time of writing, we weren’t seeing ads in SGE’s conversational mode, but this is sure to change. In conversational mode, will ads become less recognisable and harder for users to spot (e.g potentially more persuasive, for better or worse). As users move further into the conversation funnel, how will CPCs be impacted to reflect the conversation moving on? Will Google interpret this as a more engaged user, and charge more for a click? We might even see new metrics in Google Ads to reflect the “layered depth” of a keyword with indicative CPC pricing.

Meta's AI Sandbox

Meta's AI Sandbox offers intriguing opportunities for creative production. With ad text variations, background images, and cropping readily available, smaller charities with limited resources can create ads and visuals more quickly. However, it also poses challenges. AI Sandbox is not yet at a stage whereby users can simply copy and paste the outputs, meaning that setup time for campaigns still needs to be factored in. Human editing remains crucial as AI does not always capture the intended tone or messaging accurately—something that is critical for all our charities.

While Meta’s AI Sandbox is still being rolled out, we’ve been experimenting with tools like Midjourney for creating social media assets. Midjourney is a powerful tool for organisations lacking compelling assets for campaigns. Through experimentation, we’ve found it enables the creation of infographic-style creatives that we know work well to engage supporters. However, it’s worth noting that without input from a designer, some assets generated by Midjourney may be insensitive or unsuitable for charity ads. It’s also still unclear how tools like Midjourney fit into the world of copyright infringement which is something we want to keep in mind as this tool evolves, and certainly before we produce creatives for client projects.

ChatGPT for streamlining processes

ChatGPT has been useful for enhancing certain processes within our team. For example, non-developers in the team are now able to easily create and test Google Ads Scripts to improve how we monitor and manage campaigns. As always, troubleshooting is still required, as ChatGPT sometimes combines code from multiple languages, resulting in scripts that do not function as expected. Similarly, it can assist with ad copy and keyword suggestions, but human editing remains essential to ensure alignment with the organisation's tone of voice and messaging.

Sara Cox, Head of Data & Measurement

Even without the arrival of AI, the analytics landscape is changing at a rate unseen before. With a cookieless future on the horizon, UK Brexit Data Reform and the arrival of Google Analytics 4, there’s a lot to keep up with in the data and measurement space. Here are some of the areas we see as being impacted by the arrival of AI.

An efficiency driven shift to delivering greater insights

As with all fields of digital marketing, AI will bring with it efficiencies that will save time in our day-to-day processes, but this is a particularly welcome development for our data team. All data analysts encounter the same problem: most of our time is spent on data collection, leaving very little time for delivering actionable insights. Efficiencies delivered by AI that change this dynamic are really exciting. Practical, real world efficiencies might include:

  • Google Tag Manager auditing and implementation support, such as testing if a tag is returning an event, or underpinning the white labelling of a new container’s creation and tag/trigger configurations
  • Regex creation and testing for use during Google Looker Studio dashboard creation
  • Javascript programming
  • Data error correction work

Interacting with first-party data

How users interact with their own data is one exciting development that AI will almost certainly push forward. This is an area that our founder, Tom Dyson, is working on via Torchbox’s Conversations with Data (CoDa); this tool allows users to “speak” directly to first-party CRM data, which could free up time previously spent writing and executing complex SQL for the same purposes. This is not an entirely new area for our analysts: anyone who has used Universal Analytics over the last few years will have noticed the Insights functionality that has previously used machine learning to return insights for its users based on conversational interactions with GA data. How useful that functionality has been is a moot point, but the signs were there for us that AI was where Google’s analytics focus was heading.

AI as Analyst?

Articles have been published in recent months speculating about how AI might fully replace data analysts and it has made me stop and think about how likely this scenario might be. While AI can absolutely bring efficiency-centred value to data analytics there are a few things that it cannot yet do alone:

  • Context. The cornerstone of good data and insight is context. An analyst will know the right questions to ask, how to ask them and where to find the answers. They also hold all the contextual knowledge of the website or product being analysed, along with specific information about a company’s set up, structure and organisational priorities. This is the only true way to translate data into realistic actions that meet business goals.
  • Accuracy. A much reported feature of AI is the certainty with which each of the various LLMs answer questions. When they don’t know answers, they even go as far as to “hallucinate” (make things up). Therefore the role of “analyst as fact-checker” is more critical than ever. You could feed an LLM your data, ask it a question, and receive a credible solution; but a data analyst would still need to check the SQL it returns, the logic behind what approach it used and also its process for coming to a certain conclusion.
  • Communication & Stakeholder Management. LLMs can’t speak to your stakeholders to ensure there’s a voice for insight in your meetings, nor explain to you what data means in real world terms. That’s where the true value of your analyst comes in. A human will still be best placed to understand how stakeholders best process information given to them so that they action it - making reports, dashboards and presentations actually useful for your team.

To fully harness the power of AI within the analytics realm, the key is likely going to be in getting AI tools to do the manual, repetitive tasks as well as leaning on it to improve your data collection processes. There’s a lot of potential and exciting opportunities with AI that allow us as web analysts to get to the good stuff: the insight & actions.

Interested in harnessing the power of AI for your nonprofit marketing?

Get in touch for a chat and we can start exploring what's possible.