AI in practice: 5 questions accountants and bookkeepers are asking right now

Accountants and bookkeepers aren’t panicking about AI.

They’re asking sharper questions.

Recent Sage research shows that many UK firms now expect AI to benefit their practice in the next year, signalling a shift from curiosity to confident experimentation.

Across research, interviews, events, and Sage customer stories, five questions keep coming up that define how accountants and bookkeepers approach AI in practice, and reality.

That’s what we discuss in this article, as follows:

1. Where does AI make a difference?

Across accounting and bookkeeping, the first question isn’t philosophical.

It’s practical.

Where do I start? What’s worth automating?

That recent Sage research shows accountants and bookkeepers are expanding their advisory role and experimenting with new technology, often using multiple platforms.

“Leads came in from everywhere—website, socials, referrals—and no one knew what was happening,” said Jeri Williams, founder of Smooth Accounting, speaking at a keynote session at Leaders in Accounting.

Once we used AI tools to automate and summarise calls, everyone could see what was going on. Conversions shot up.

Jeri Williams, Smooth Accounting

Jeri’s story mirrors many others.

Firms are finding wins in automating repetitive, data-heavy workflows, such as onboarding, reconciliation, and disclosure checks, while keeping humans in control of review and advice.

This is agentic AI in action: systems that operate on behalf of humans but always within defined boundaries.

Accountants set the intent; AI executes the repeatable steps. The professional judgment calls, and the “should we?”, not the “how do we?” checks, stay firmly human.

Across UK SMBs, Sage Copilot is automating busywork like drafting emails, chasing payments, and summarising activity, so teams can focus on judgement calls.

At Watson’s Anodising, for example, Copilot runs in the background, freeing hours each week for production and customer service.

2. How do I stay compliant and trusted?

Trust, reliability, and compliance are the top priorities for accountants when recommending software, further Sage research reveals.

While speed and efficiency are valuable, they mean little without a foundation of trust and clear accountability for AI-generated outputs.

Initiatives like the Sage Trust Label exemplify this approach, showing exactly how AI features are designed, what data they use, and how security is maintained.

When barriers to adoption often stem from unfamiliarity and perceived complexity, the key to overcoming them is to highlight the importance of transparent AI policies and robust human oversight.

Forward-thinking firms are formalising these principles into lightweight digital codes of conduct, creating an easily digestible way to build client confidence.

AI isn’t magic. Be crystal-clear where it helps and where you still need human oversight. If expectations are unrealistic, disappointment follows—and people stop using it.

Mohammed Sidat, associate director, Wolters Kluwer

What a simple AI policy should cover

Here are some suggestions of what should be included:

Purpose & scope

Set clear intent for using AI in your firm.

Define approved use cases, such as data analysis, draft preparation, or summarisation.

Establish boundaries so everyone understands what’s in scope, and what isn’t.

Privacy & data handling

Clarify how information moves through your systems and where it’s stored, showing how you comply with local and regional data and privacy laws.

Use only vetted tools and never upload client data to public models.

Maintain a list of approved platforms with data-protection notes.

Accountability & human review

Specify who signs off on AI work and ensure human oversight for all client-facing content.

Show clients there are humans in control. The professional using AI remains responsible for accuracy.

Compliance & ethics

Anchor your policy in existing professional standards, including copyright guidance and record-keeping obligations.

Show your tools stay up to date with changing regulations.

Treat AI prompts and outputs like any other working paper—store data safely, provide version-control, and review them often.

Consent & transparency

Show clients how AI fits into your workflow so they can understand where to expect it and how it helps them.

At Smooth Accounting, Jeri Williams added a simple consent line to all meeting bookings after clients questioned the extra “guest” on their calls.

“We just explain it’s our AI note-taker, and they’ve already agreed when they booked. It’s become second nature.”

A one-page policy is quick to draft and saves hours of uncertainty, strengthening client confidence in your commitment to privacy and professional responsibility.

These principles aren’t theoretical. They deliver real results in practice.

Sage customer British Veteran Owned (BVO) reports saving £15,000 and 36 days per year with Sage automation, all while strengthening client confidence through a transparent process.

3. What does this mean for my role?

As AI becomes an integral part of daily operations, your AI-inclusivity and use visibility, is increasingly viewed as a competitive advantage.

The biggest question for accountants and bookkeepers is personal: if AI can handle the admin, what’s next on the list? The answer is up to you, as you’ll have more time to prioritise client care and relationships, or simply finish work on time.

AI gets you 90 percent of the way. That last 10 percent—the context, the empathy, the credibility – is where we earn our value.

Jared Goodrich, director of innovation, Moore Kingston Smith

Walter Dawson & Son, a Yorkshire-based firm, demonstrates how embracing Sage and AutoEntry has freed their team from manual admin and helped them prepare early for the shift toward Making Tax Digital for Income Tax.

The result is more time for client-facing advisory work rather than form-filling and data chasing.

Managing Director Julie Young notes in the success story: “The time saving from Sage and AutoEntry is used back with the clients. It allows us to reflect on the information and talk to our clients more about how they’re performing.”

Walter Dawson & Son accountant Hannah Hall, adds, “Sage Copilot is going to allow us more time to go out and see our clients, listen to their needs, and be more people-facing rather than behind the computer.”

Together, their experience shows how AI and automation free up time for deeper client engagement and strategic advice.

As roles evolve, the opportunity is to move beyond simple automation into tools that amplify your judgment, where you set the direction and technology accelerates the work.

4. How do I go from automation to agency?

You’ve seen that automation saves time.

You know that agency creates value.

Agentic AI connects the two, giving you systems that can act independently, but always under your instruction and with your judgment in control.

This is the shift that turns tools into teammates, moving from assistance to action while keeping the professional firmly at the wheel.

Think of Sage Copilot: it draws from accounting data to draft, suggest, and surface insights, but the practitioner remains at the controls.

You decide the goal. Copilot handles the legwork.

The same principle applies to MTD submissions, reconciliations, or reporting workflows. When professionals set the intent and validate the output, AI becomes an extension of expertise, not a substitute for it.

The new Sage MTD Agent, for example, simplifies compliance by automatically preparing and submitting updates to HMRC while flagging exceptions for review.

This practical example of Agentic AI is built for real-world accounting. It offers automation that works within defined professional boundaries and strengthens human oversight rather than replacing it.

Other Sage Ai Agents—including the new Finance Intelligence, Close, and Assurance Agents—are built to operate autonomously within clear, trusted boundaries. They focus on gathering data, generating summaries, or reconciling figures, then presenting the results for human review.

Agentic AI builds on the foundation laid by Making Tax Digital—moving from simply connecting systems to pre-emptive action, so your software anticipates what should happen next while you stay in control.

For a deeper dive into how this shift is transforming accounting workflows, read our companion piece Authentic intelligence in action: How Agentic AI will shape the future of accounting.

5. How do I start?

Accountants and bookkeepers often say the same thing: the hardest part of AI isn’t using it—it’s starting.

Ease of use, familiarity, and value for money are key factors in sifting through software recommendations, while cost and feature limitations remain common obstacles.

You don’t need a full-scale strategy or a new tech stack—just try one contained experiment to prove what’s possible.

These agentic principles also apply at firm level: start small, act with intent, learn fast.

“Progress beats perfection,” said Mohammed Sidat, speaking in a keynote at Leaders in Accounting. “Try one repetitive task, learn from it, then repeat. You’ll discover what works for your firm instead of waiting for the perfect solution.”

Think of it as your 30-day AI sprint—a safe, structured way to see what human-plus-machine collaboration looks like in your own workflow:

Week 1: Pick one repetitive workflow

Choose one repetitive, measurable workflow—like onboarding, invoice chasing, or disclosure checks—and record your current time or cost baseline.

Week 2: Pilot an automation or AI feature

Test an AI feature or automation using tools you already trust.

As Jeri Williams put it, “Half the time, you’re already paying for AI. You just haven’t switched it on yet.”

Week 3: Add review and consent gates

Decide where human sign-off is required and how client consent is captured.

Week 4: Measure the impact and share the learning

Track the impact—hours saved, quality improved, errors reduced—and share results internally to encourage further experimentation.

Each small experiment builds your firm’s collective intelligence and demonstrates how, by setting intent and AI executing safely, you can deliver real client value.

Final thoughts: The real future of AI in accounting

Accountants who understand AI—its boundaries, biases, and potential—will shape the future of the profession. So, by just getting started, you’re already halfway there.

Throughout history, innovations like double-entry bookkeeping, spreadsheets, and cloud technology have expanded accountants’ agency.

More recently, initiatives such as Making Tax Digital have accelerated the shift to digital workflows, giving accountants greater control and insight over their work.

Agentic AI is the next chapter: where software is your teammate, assistant, or coach, not simply a passive tool.

Accountants and bookkeepers have always led through change; now, AI offers a new way to demonstrate that leadership to your firm and your clients.