Volume 1 Edition 7 | by DATAFORT Founder Marcie D. Terman | 21 September 2026
The clock is ticking, and the next Making Tax Digital (MTD) quarterly deadline is less than seven weeks away, on 7 November.
The August cycle gave practices their first real look at where the additional workload actually appears: which clients require constant follow-up, where records arrive incomplete, where staff have to intervene manually, and where minor administrative tasks quietly stack up and multiply.
That workload will only grow. MTD for Income Tax currently applies to sole traders and landlords with qualifying income above £50,000. In April 2027, the threshold drops to £30,000, bringing a substantial new cohort of clients into scope, followed by a £20,000 threshold in April 2028.
The critical question for practice leaders isn’t simply: How do we survive the next deadline?
It’s: If our processes are struggling today, what will they look like in April 2027?
BE INTENTIONAL IN YOUR RESPONSE TO ADDED WORKLOAD
When workload expands, the immediate instinct is often to hire more staff. But today, recruitment isn’t the only option available to accountancy teams. Modern technology allows existing teams to handle many current workflows far more efficiently.
Before bringing in developers or evaluating new software, ask one fundamental question: Do you believe your current process is as good as it should be?
One of the easiest ways to waste money on technology is to take a flawed process and make it run faster. An unnecessary hand-off remains an unnecessary hand-off. Re-entering data into multiple disconnected systems simply happens more quickly. A badly designed chasing workflow will generate reminder emails automatically, potentially annoying clients faster and more frequently.
You may have automated a task, but you haven’t necessarily improved the business.
A TYPICAL MTD EXAMPLE
Consider something as ordinary as collecting information from a client.
The practice requests bank statements for June, July and August. Records arrive piecemeal by email; staff open messages, download attachments, file them and update practice software. A week later, someone notices that June’s statement is missing. Chasing begins and, as the deadline approaches, senior staff may get pulled in because the return is blocked.
None of these individual actions looks onerous in isolation. Multiplied across dozens or hundreds of clients, however, they represent a substantial administrative burden.
If you simply use AI to write and send those reminder emails, you save a few minutes, but you have automated only the most visible part of a broken workflow.
WHAT IF WE REDESIGNED THE PROCESS FIRST?
A better starting point is to ask what the workflow is actually trying to achieve. The practice needs to know what was requested, recognise what has arrived, identify missing items immediately, prompt the client for anything outstanding and stop chasing as soon as the information arrives and has been checked for readability. Genuine exceptions should then be routed to human staff only when judgement or intervention is required.
Once defined this way, the solution looks completely different.
A properly designed workflow can track incoming files, prompt the client for June’s statement without a member of staff first having to spot the gap, check that the document is readable and stop asking once it has been received.
If a document is corrupt, unreadable or contains conflicting information, it can be flagged directly to the appropriate member of the accountancy team. Depending on how the process is designed, the system can notify the client, the team, or both.
The human doesn’t disappear; the human stops doing work that no longer requires human intervention.
UNCOVERING HIDDEN INEFFICIENCIES
Preparing a business process for automation forces you to describe it step by step. The moment you do, long-standing inefficiencies become easier to see.
Why is data copied manually between systems instead of being synchronised? Why is a qualified professional checking routine document arrivals? Why do redundant internal sign-off stages exist? Why maintain side spreadsheets when the same information already exists in practice software?
An automation project can act as a catalyst for business improvement — not because AI inherently knows how your practice should run, but because preparing for automation forces you to question assumptions that may have been buried in the process for years.
AUTOMATION CAN ALSO INHERIT BAD CONTROLS
Process redesign is not only about removing wasted effort. It can also expose controls that should have been there all along.
Take a simple refund process. Imagine that the system allows a refund to be issued without first checking that the amount being refunded does not exceed the value of the original purchase. That is already a poorly controlled process. A human operating within it could potentially exploit the weakness by issuing an excessive refund, perhaps directing the money to an account they control.
Now give an AI agent authority to execute exactly the same process.
If nobody has built in a rule tying the refund to the original transaction, the underlying weakness remains. The agent could potentially process a refund greater than the value of the purchase because the process itself contains no control preventing it.
The problem is not that AI has suddenly invented a new type of fraud. The weakness already existed. Automation simply inherited it.
The correct response is to design the control into the process: the refund should be matched to the original transaction, appropriate limits should be enforced automatically, and anything outside those limits should stop and require human approval.
The same principle applies to other financial processes. If changing a supplier’s bank details and authorising the resulting payment should never be controlled by one person without an independent check, an automated system should not be given that authority simply because it is technically capable of performing both actions.
That is why process mapping before automation is about more than efficiency. It can expose weak controls as well as wasted effort.
SHIFTING TIME TO HIGH-VALUE WORK
The true cost of operational friction isn’t just lost minutes. It is how those minutes could have been spent more productively.
When qualified accounting professionals spend time tracking missing documents, updating statuses or re-keying data, practice margins can erode. The goal of process redesign isn’t to squeeze every second out of the working day. It is to redirect professional time towards work clients actually value: resolving complex tax exceptions, interpreting business metrics, providing proactive advisory services, and meeting prospective clients or developing new relationships.
When routine administration is reduced, capacity can increase naturally. Revenue may then be able to grow without employment costs increasing at exactly the same rate.
That is where the economics of automation become more interesting. The benefit is not simply that one administrative task happens more quickly. It is that professional attention can be moved towards the work where professional judgement creates value.
START WITH THE PROCESS, NOT THE AI
When planning your operational strategy, stick to a clear sequence:
MAP THE PROCESS → IDENTIFY FRICTION AND RISK → QUESTION EVERY STEP → REDESIGN THE WORKFLOW WITH THE RIGHT CONTROLS → AUTOMATE
Avoid the shortcut:
FLAWED PROCESS + AI = FASTER FLAWED PROCESS
Instead of asking where an AI tool can fit into your existing workflow, turn the question around:
“If we were designing this process today, with modern tools available, would we design it this way at all?”
That question matters because automation does not distinguish automatically between the useful and useless parts of an existing process. If you reproduce unnecessary work, it can reproduce unnecessary work. If you reproduce weak controls, it can reproduce weak controls.
The objective is not maximum automation.
The objective is a better process.
WANT TO FIND WHERE AI CAN SAVE TIME IN YOUR PRACTICE?
If a core workflow is consuming too much team time or creating friction, DATAFORT can help you audit, redesign and structure it before any technology is implemented.
Every month, we offer four complimentary half-day consultancy sessions for practice leaders. We will map one time-intensive process in your business, uncover hidden bottlenecks, risks and redundant steps, and build a clear roadmap for process optimisation and automation.
Applications are open for October’s allocation. Book a 20-minute discovery call with our team to bring one process for discussion and find out whether your practice qualifies.
NEXT EDITION: YOUR PRACTICE ALREADY HAS THE DATA — ARE YOU LEARNING FROM IT?
Every late client, repeated request, missed deadline and piece of rework leaves a trail of data behind. Most practices use that information to complete the immediate job and move on.
But looked at across hundreds or thousands of transactions, those same records can reveal where clients struggle, where staff time disappears, which processes repeatedly cause problems and where margins may be quietly eroded.
In the next edition, we’ll look at how AI can analyse patterns in the operational data your practice is already producing — and turn that information into better decisions about how the business is run.
OFFICIAL SOURCES & GUIDANCE
- HMRC: Use Making Tax Digital for Income Tax
- HMRC: Quarterly updates for Making Tax Digital
- HMRC: Get ready for MTD: Agent Toolkit
- MTD: Burden or Business Advantage, our handy e-guide on how AI could streamline the peaks and valleys of work that have resulted from the introduction of the Making Tax Digital initiative.
