• Register
Return to: Home > Comments > Artificial intelligence: a smooth adoption of automation

Artificial intelligence: a smooth adoption of automation

It is one of the critical topics of the day, but what steps can accounting firms take to adopt AI into their practices? Will Shilson, VP of client technology at Calligo, provides a practical overview.

While the AI and automation conversation has dominated the accountancy space for a while now, it is clear there remains a widespread lack of clarity about how to actually implement these new technologies.

The principal concern is not a matter of how, or even where to find the skills required: it is more commonly a matter of exactly where they are needed. But this is no easier a question to answer. Where does a business start in trying to identify where AI or automation can make the most difference? 

The best route for any accountancy firm is to take a step back. Do not focus on the technology, or even its promise: focus instead on your processes and your data, and then on identifying where the most value-adding improvements can be gained. By obsessing over a pre-selected technology and where to put it, you will often force a solution into a problem that it does not truly solve, or that does not need solving as urgently as another.

For many accounting firms, we find the greatest benefits are gained from identifying the processes that drain the most resource time, are most repetitive and most formulaic – and probably also demoralise consultants the most.

You will probably find that these are comparatively simple matters of automation rather than the more adventurous deep insight projects – for example, automating the information-gathering processes surrounding personal tax. They also tend to be the projects that yield the most immediately important gains sooner, and can act as springboards into more dramatic projects further down the line.

Here is a guide to how accounting firms can take a more informed and practical approach to smooth the adoption of AI and automation.

Understand where the value lies

Investigate which processes are most suitable for automation, have the greatest potential value, and have the least risk to the business.

Like any discovery phase, the principle is to ask questions, leaving nothing to assumption.

AI and automation depend on data, so the key to this phase is understanding how data flows into, around and out of your business – and where the bottlenecks and most resource-intensive places are. 

Maturity assessment

The business needs sufficient technical self-awareness and foundations to make the most of its data and to automate a business process. AI and automation are independent data-processing tasks; however, they still require the appropriate technical architecture to launch from and the skills to maintain it before they can be rolled out.

Beyond technology, automation also needs a suitably pro-innovation culture to nurture it. One of the key questions will be whether you see data as being integral to your overall business strategy. Do you see it as a tool or an asset? Do you base your key business decisions on data? 

This stage can be the longest of the entire process, as most businesses will have some extensive work to do here to prepare themselves for the introduction of such innovation.


Once you have ensured that your business is suitably data-mature to embark on the project, create a roadmap that identifies its objectives, acknowledges the obstacles, and shows at a high level how they will be overcome and what results are expected.

Look wider than the project itself, and include how it can integrate with or even impair wider initiatives and plan accordingly. 

Solution architecture

How is it going to work? Consider the practicalities of the project, the technical processes and requirements, the responsibilities and most importantly, the privacy of any data subjects whose data may be used.

Privacy by design should be a prerequisite for any project that uses data – not just because legislation such as GDPR demands it, but because it is the ethical way to treat data.

Ensure all stakeholders are informed

This is one of the most delicate parts of the process. It is where reality hits and where any concerns are aired.

Stakeholders need to know what they are signing up for, what the consequences to other initiatives may be, the costs and ROI, the likelihood of success and how the project fits within the wider strategic initiatives.

Try a prototype or soft launch

This is the ideal time to finesse what you will measure to prove eventual success or failure, to conduct any final checks of data sources and their suitability, and test the reliability and accuracy of any integrations.

Integration and review

Once you are happy with how it is running and that the potential benefits remain likely, set the process live and ensure you are monitoring it and anything related to it closely in the first few weeks.

Data is being collected, algorithms are learning from outcomes and performance is improving every day, but has the project done what it set out to do? Is it delivering the cost and efficiency savings planned, or the insights expected? Simply, is it successful?

Review this area regularly, and to ensure that you are meeting the ROI set out in your initial proposals. 

Top Content

    ARGA team, assemble!

    The new top team has been named that will see in root-and-branch reform at the Financial Reporting Council (FRC) as it transforms into the Audit, Reporting and Governance Authority (ARGA). Will the new duo be as dynamic as some are hoping? Robin Amlôt reports.

    read more

    FASB: a quest for simpler standards

    FASB chair Russell Golden addressed the IMA 2019 Annual Conference and Expo at the Sheraton San Diego Hotel and Marina, California, on 18 June. IMA immediate former chair-emeritus Alex Eng acted as moderator. Joe Pickard reports.

    read more

    The future of audit, and how to get there

    Two recent reports peer into the future of the audit profession. One analyses what an audit should offer, while the other looks at how the audit process will be carried out. Robin Amlôt takes a closer look at both.

    read more

    EFAA elects new president, focuses on digital future

    EFAA’s new president, Salvador Marin, outlined his key priorities for the next two years at the organisation’s 2019 annual general meeting, while outgoing president Bodo Richardt offered advice. Robin Amlôt reports.

    read more


    As the Coronavirus (COVID-19) continues to spread across the world, the International Accounting Bulletin and The Accountant will be collating all the latest news and updates from the profession on the pandemic’s impact.

    read more
Privacy Policy

We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.