We partner with GoProposal by Sage. I, Jo, used GoProposal in my business, and it revolutionised the way I priced my services. And the next thing that’s going to revolutionise your practice is AI. So when Sage invited us to attend, I couldn’t say no, and I dragged Zoe along too.
This article contains paid promotion for our partner GoProposal by Sage
I came away thinking the conversation around AI in bookkeeping is finally getting more useful.
And that’s important, because I think a lot of us are tired of hearing two extremes. On one side, AI is presented as the answer to everything. On the other, it is treated as something to be feared, avoided or dismissed. What I heard in Bristol was this.
AI is not a replacement for professional judgement, and it is not a magic shortcut either. But it is becoming increasingly practical inside bookkeeping and accounting workflows, and the firms that benefit most will be the ones that approach it curiousity – and control.
That felt like a much better conversation.
One of the strongest themes from the day was that AI in practice is not arriving as one big shift. It is showing up in lots of smaller ways already. Meeting notes. Transcriptions. Draft emails. Document processing. Summaries. Categorisation. Workflow support. Research assistance. Review prompts. The kind of work that sits around the edges of your day, slows everything down, and eats capacity without always feeling “important” enough to fix.
That is part of what made the Bristol session interesting. The reality is that many firms are already using AI in some form, whether deliberately or unofficially. The research discussed on the day referenced a survey of 600 practitioners across the accounting profession, with 80% reporting at least a minimal benefit from AI technologies, and around two thirds already encountering AI within their core accounting software.
We’re seeing this inside The Bookkeepers’ Collective too, and it’s so important to us that bookkeepers are aware of and embracing new technology that we built a whole collection of training around it in The AI Collection.
The big opportunity is not in chasing every new shiny tool, and we all know that. But there are huge opportunities in the AI that is being embedded into the software that we already use and trust. This is incredibly important for me because as bookkeepers, trust, accuracy and data security matter above everything else.
My second biggest takeaway was that the conversation has matured.
A couple of years ago, most AI discussions in our world seemed to start and end with, “Will this replace us?” Now the better question is, “Where does this help, where does it not help, and what guardrails do we need?”
There was a lot of emphasis in Bristol on the fact that accountants and bookkeepers are custodians of highly sensitive client information. So yes, efficiency matters. But security matters more. Governance matters more. Oversight matters more. The challenge is not simply getting AI to do something quickly. The challenge is using it in a way that protects clients, and protects the reputation of the practice using it.
That is why one of the most useful points discussed was this idea that AI needs to be treated like a team member.
If you brought a junior into your firm, you would not hand over access to everything, give no guidance, skip the review process and hope for the best. You would train them, document expectations, set boundaries, review outputs and build confidence over time. The same logic applies here. AI can help with the draft. It can help with the process. It can help with the first pass. But responsibility still sits with the bookkeeper, so human judgement still matters.
I thought this point was explained really well on the day: the difference between an assistant and an agent.
An AI assistant helps with a task. An AI agent helps move an entire workflow forward.
I see a lot of bookkeepers thinking in terms of “Can AI write an email?” when the question they could be asking is “Can AI support the whole process around that email?” For example: gather context, summarise notes, draft a response, flag missing information, suggest next steps, and leave the final judgement to the human reviewing it. That is a much more practical way to think about where this technology is going.
But for me, the most important part of the Bristol roadshow was not the technology at all.
It was the reminder that the end goal is not to become less human.
It is to create more room for the human parts of the job.
That is the bit I think our profession needs to hold onto.
Because the value of a bookkeeper has never been in typing numbers into software for the sake of it. The value is judgement, pattern recognition, asking the right questions, giving reassurance, catching the thing a client has missed, asking another question, and another, and a better one. With the ultimate goal of helping a business owner make a better decision. AI may well reduce some of the manual work, but that does not reduce our value. If anything, it should give us more space to deliver the parts of the job clients actually remember and pay for.
That links to pricing too, which is where the conversation gets really interesting.
If AI helps you do parts of the work faster, does that mean your prices should go down?
No.
And that seemed to be one of the underlying themes of the day as well. Faster is not the value. Better is the value. More insight is the value. Better service is the value. More capacity, fewer bottlenecks, stronger consistency and a better client experience are the value.
But that does mean that pricing might need to be different in the future.
We moved from hourly pricing to transaction pricing.
We teach our community fixed pricing for what we call a “gold standard service”.
But last year, Zoe presented an idea that I don’t think the audience was ready for.
A performance fee.
She called it a base fee + commission on performance.
What if bookkeepers were paid for helping their clients create more profitable businesses?
What if fees were tied to the client hitting a financial goal?
And this isn’t as out there as you might think. We already see this in the R&D claim space.
If AI allows a practice to reduce drudge work and improve the financial outcomes for clients – which it absolutely can do when AI is plugged into the figures, then the commercial question is not “How cheaply can we now do this?” but “How do we price in a way that reflects the outcome we are creating?”
That is also why I liked that the roadshow felt like a working session about how firms can sensibly build businesses for the future.
And you can start small. Look at the tasks that create the most friction. Identify repeatable workflows. Pay attention to where data lives. Decide what should never be copied and pasted into random tools. Put some governance in place. Train your team. Experiment in a controlled way. Review what works. Then build from there.
Honestly, that is probably the biggest lesson I took away from Bristol: the firms who will get the most from AI are unlikely to be the firms doing the most talking about AI. They will be the firms building better workflows, better standards and better habits around it.
Thoughtful implementation.
For bookkeepers especially, that feels like the right approach. We are already managing deadline pressure, workflow complexity, client expectations and capacity constraints. The answer is not to throw more tools into the business. The answer is to reduce friction in the right places and protect client trust.
Reading about AI is one thing. Hearing hot takes on social media is another. But sitting in a room with other practitioners and seeing a practical discussion about what adoption could actually look like inside a firm is something we all need to be doing, regularly.
There are more roadshows coming up.
And, would I recommend going?
Yes. Especially if you want a more balanced view of where AI fits in bookkeeping and accounting right now.
I don’t think every practice has to or wants to become AI-first.
But I do think every practice does need to become more intentional about how work gets done, where time is being lost, and how to protect capacity as the profession changes.
If you missed Bristol, there are still two Sage AI Roadshows to come:
Manchester on 16 April 2026 and London on 28 April 2026. These are worth attending if you want a realistic, hands-on look at how AI can support your workflows without replacing the judgement, trust and human insight your clients rely on.
And that, for me, was the real takeaway from the day.
The future is not about becoming less human.
It is about using the right tools so we can be more useful, more thoughtful and more present where it counts.
Sign up for the roadshows, here.


