The AA (Automobile Association) is one of the UK’s largest roadside assistance and motoring services providers with over 14 million members. It had been experimenting with generative AI for twelve months, but without a coherent model for adoption, application, or measurement. Recently, however, it has rolled ChatGPT Enterprise out to 1,500 people.\u00a0<\/p>\n
This change in gears was enabled by OpenAI partner Cognita Reply, which is a vertical startup within the IT consultancy Reply Group. \u00a0Karen Tew, Senior Manager at Cognita Reply, specializes in devising workplace training structures and began working in the AI area with CoPilot a couple of years ago prior to focusing on ChatGPT at Cognita Reply.\u00a0<\/p>\n
Other parts of the Reply Group had worked with the AA and were already involved with them around AI, helping them create an internal narrative around their goals within the business. OpenAI was also talking to the AA at the same time, so when Tew arrived the company already had a few trial licenses, but usage was disparate – some areas were quite advanced such as the engineering department but other areas had never used AI. Tew explains:<\/p>\n
I don\u2019t look at organizational maturity because you need to look at the right joining point for everybody, so we provide different levels of training starting at level zero, which takes a \u2018no-one left behind\u2019 philosophy and then you have also got those who are already dabbling with Claude. With the AA, the majority of people attended Level Zero training and I wasn\u2019t expecting this but we provided people with the option to drop it, if it was repeating things they already knew.\u00a0<\/p>\n
The viewpoint for training has to be that this thing (AI) is not going away and you are going to have to use it, so let\u2019s ground it in your reality and make it relevant for your job. As an employee you need to think about what you can use it for as well as what you should use it for.<\/p>\n
The key to overcoming a lot of fears is to choose the right initial use cases that are carefully defined but to also create an environment where people can get it wrong sometimes. You also, of course, need the right guardrails and validation steps in place.<\/p>\n
Putting the mechanism to innovate in employee hands\u00a0<\/p>\n
The training that Cognita Reply provided featured a hackathon-style session for each employee to identify repetitive work where they feel that they are a blocker or are creating a bottleneck. Tew continues:\u00a0<\/p>\n
Prior to the hackathon we provided a custom GPT training session and sent out some pre-reads explaining the level of knowledge that would be gained by taking the challenge. We got so many submissions that we had to use qualification criteria to filter a certain number through while also providing feedback to each applicant as to what would be required prior to the build phase in terms of the details of the business process and its intricacies. Within the hackathon, we look at parts of that process using the knowledge and experience of the staff in that area and help them apply the technology. The value is all in the nuance that the employee has and the consultant can rarely provide that.<\/p>\n
Tew advises deploying the most universal use case as the best starting point to get the whole organisation motivated and involved. Then trainers need to be working closely with the teams that work in control functions such as Risk, Legal and so on as well as with the IT teams for them to understand what they are enabling for users. Finally, an enablement program should bring in functional department leads. With these four perspectives in place organisations are in a position to drive use cases.<\/p>\n
Tew is keen to point out how empowering this technology can be for workers that are not used to having their ideas heard. She advises:\u00a0<\/p>\n
A soft benefit which is often over-looked is that this technology empowers people to innovate because once they understand how to use AI people feel more comfortable coming forward with ideas to innovate. These are often people that had no language or mechanism to do this before. This is all part of the new way of working because it can enable each person in the organisation to free up 10% of their time to innovate. The saving comes from not having to respond to generic questions or undertake repetitive tasks.<\/p>\n
Cognita Reply did make use of community champions in the AA training by inviting people to volunteer as well as getting managers to nominate people. Consulting support was made available for all these active champions with consultants available to jump on five-minute phone calls as required. Tew says:\u00a0<\/p>\n
One aspect of this is that this technology is not a once and done training proposition because every four to six weeks a new model is released. Organizations need the foundations in place to stay on top of this, to know what is emerging and how it can be made useful for the organization. Champions are necessary to keep this process going.<\/p>\n
As a measurement of success, active use among licence holders at the AA went from 29% to 63% between December 2025 and March 2026. Tew adds:\u00a0<\/p>\n
In terms of metrics, we are monitoring creative and repeat uses, as well as return uses so that teams that feel disconnected do not get left behind. You have to be careful your stats are not being skewed by the early adopters, and you have to track what features are being used: GPT, Chats, Codex and how many tokens are being consumed across different functions.<\/p>\n
My take<\/p>\n
When it comes to training for adoption of AI, Tew\u2019s work demonstrates the need to get inside the mindset of different types of users and ground experimentation in what people actually do on a daily basis in tasks that they can articulate. As Ben Bradford, Director of AI at the AA told Cognita Reply:\u00a0<\/p>\n
You can\u2019t really teach anybody how to use these tools for their particular job \u2013 the nature of them is they\u2019re so multi-faceted, and everyone\u2019s job is so different. What you can do is teach a core set of personal capabilities and give people the confidence and freedom to go away and experiment. That\u2019s what this program has done.<\/p>\n","protected":false},"excerpt":{"rendered":"The AA (Automobile Association) is one of the UK’s largest roadside assistance and motoring services providers with over…\n","protected":false},"author":2,"featured_media":142023,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[7],"tags":[521,580,10885,23324,157,14652],"class_list":["post-142022","post","type-post","status-publish","format-standard","has-post-thumbnail","category-openai","tag-ai-adoption","tag-chatgpt","tag-data-management","tag-generative-ai-and-llms","tag-openai","tag-use-cases"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/142022","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=142022"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/142022\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/142023"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=142022"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=142022"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=142022"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}