How bookkeepers are the human API powering UK small businesses

At the Institute of Chartered Bookkeepers (ICB) Bookkeepers Summit, CEO Amy Copeland described bookkeepers as human Application Programming Interfaces (API): the connective layer linking the messy, mismatched systems that small businesses depend on but rarely understand.

Amy said, “Bookkeepers have become the API. The thing that connects two bits of software so that data can flow. And we’re saying bookkeepers are doing that for everything that matters”

And she’s right.

“Human API” is a great description of the work bookkeepers do every day. In this article we explore the concept. Here’s what we discuss.

The human API: The invisible infrastructure small businesses run on

As bookkeepers, you’ll carry out tasks like:

• Translating HMRC’s digital vision into something clients can act on.
• Connecting VAT deadlines with a sole trader’s forgotten Government Gateway login.
• Catching cash flow danger before the business owner even realises their supplier raised prices six months ago.

You sit at the intersection of data, behaviour, compliance, and chaos—and you make an ecosystem work that could collapse without you.

If any part of that ecosystem breaks, it lands on you first.

And when it works, it’s usually because you’ve held the pieces together long enough for everyone else to catch up.

This is what “human API” really means. It’s the real work happening every day in Britain’s small businesses.

Why “human API” matters right now

Small businesses are under pressure from every direction, and bookkeepers field almost all of it.

For example, late payments are a systemic threat.

In the UK, an estimated 38 small businesses go under every single day because cash flow dries up. Bookkeepers see those early warning signs before anyone else.

This number comes from UK Small Business Commissioner research, published alongside the government’s Small Business Plan and commissioned by the Department for Business & Trade (DBT):

The research also says:

• Late payments alone costs the UK economy £11 billion a year.
• Small-business owners spend, on average, 86 hours chasing debt per year. These are hours that produce nothing but stress and lost momentum.

So, who helps businesses face the consequences when the system breaks?

• Not policymakers.
• Not software vendors.
• Not even accountants.

It’s bookkeepers.

• You’re the ones who get the panicked emails when invoices go unpaid.
• You’re the ones who explain why the supplier raised prices or why the tax pot is too low.
• You’re the ones who quietly fix the financial admin that business owners either avoid or don’t understand.

This is why the human API metaphor is a structural reality.

And that’s why it was so significant to listen to Emma Jones, the UK Small Business Commissioner, speaking at the Bookkeepers Summit main keynote.

I simply could not have done it without my bookkeeper. She became the finance pillar of the whole business.

Emma Jones, UK Small Business Commissioner

Emma’s message was clear: bookkeepers aren’t a peripheral part of the small-business ecosystem. They are the people holding it together when cash flow breaks, when admin overwhelms owners, and when payment delays threaten survival.

At the Regional Investment Summit in October 2025, the Chancellor announced the government wants to reduce the admin burden on small businesses by 200 hours.

That only happens if bookkeepers can eliminate friction in the systems they’re forced to mediate, such as MTD ambiguity, inconsistent data, late payments, client misunderstandings, and endless reconciliations between tools that don’t communicate with each other.

In other words:

• Small businesses aren’t drowning in complexity because they lack software.
• They’re drowning because everything around them is fragmented.
• Bookkeepers keep it flowing.

At a moment when small business confidence is fragile, late payments are rising, and digital adoption is being pushed harder than ever, your human API role makes you part of the critical infrastructure in the small-business economy.

A human API case study from the UK small business commissioner

Emma Jones further shared a story that captures the human API role in practice.

Before taking on her current post, she ran several businesses and did what many founders do: kept the books herself, cross-checking bank balances and chasing invoices between emails. It worked until growth raised the stakes and the risks became real.

A mentor gave her the advice every bookkeeper already understands: Bring in a bookkeeper before something breaks.

So, she hired one part-time. That person quietly became the financial backbone of the business.

They managed credit control, cleaned data, kept everything compliant, and eventually steered the company through two major transactions: an investment round and later, a sale to a US private equity firm.

This is the human API in action. The bookkeeper didn’t just “do the books.”

They connected the operational reality, the compliance demands, the investor expectations, the data, the decisions, and the cash flow—linking parts of the business that never speak to each other unless a human steps in.

Emma’s story makes one thing unmistakably clear: a business can go a long way without an accountant, but not without a bookkeeper acting as the human API.

5 key insights from the Bookkeepers Summit

Bookkeepers sit at the centre of how small businesses experience technology, regulation, and day-to-day financial pressure.

Across the keynotes, panels, and practitioner discussions of the Bookkeepers Summit, a consistent pattern emerged: Bookkeepers operate at a different altitude, carry a different kind of load, and shape the success or failure of digital change in overlooked ways.

The Summit didn’t just celebrate the profession. It revealed how deeply bookkeepers shape the flow of data, compliance, and decisions inside small businesses.

The following themes stood out.

1. Bookkeepers operate at a different altitude

The Bookkeepers Summit made the contrast with accountants unmistakable.

Bookkeepers operate in the day-to-day flow of a business, not in quarterly review cycles. You see the micro-signals that determine whether a small business is stable or sliding, such as:

• Cash-flow tightening in real time
• Supplier price creep
• Client behaviour changes
• Early signs of disorganisation or burnout.

These are the things that never show up in a set of monthly statements until it’s too late.

Emma Jones’ opening story captured this perfectly.

The first person who stabilises a growing business isn’t the accountant doing compliance work. It’s the bookkeeper who acts as the connective layer between the owner, operations, and finances.

Bookkeepers aren’t “junior accountants”. You do a different job entirely.

• You’re inside the workflow every day.
• You see numbers as they form, not after the fact.
• You make judgment calls in the moment because you’re close to what’s happening—who’s paid, who hasn’t, what’s missing, what’s unusual.

That proximity is what makes bookkeepers the connective tissue of small businesses.

It also explains why your needs around technology and AI aren’t the same as those for accountants. Your value must show up in the daily flow, not in end-of-month tasks.

2. Bookkeepers are being forced into the regulatory translation role

Between HMRC’s digital ambition and the lived reality of MTD, bookkeepers are the ones bridging the gap.

Amy Copeland said it outright: bookkeepers connect HMRC’s “amazing vision” with the fact that “sometimes you still have to write to HMRC for an agent code.”

Every question clients ask about landlords, quarterly updates, digital records, penalties, or what’s changing versus staying the same, lands with bookkeepers first, not HMRC. It’s unpaid translation work, and it can be relentless.

This is part of the human API load and a unique skillset you must develop, turning regulatory noise into something a real person can act on.

3. MTD: Digital change succeeds or fails at the bookkeeping layer

In a Making Tax Digital update at the Bookkeepers Summit, HMRC’s Director of MTD Craig Ogilvie was blunt about the scale of what’s coming.

Craig called MTD for Income Tax “the biggest thing I’ve ever done”. He explained that it was bigger than the furlough scheme or Brexit systems—and said the real change isn’t just technical. It’s “societal”. Bookkeepers, agents, and accountants will have to work differently to the old Self Assessment norm.

“We’ve re-platformed our tax system and built new APIs so the software market can operate properly… But MTD is in safe hands with bookkeepers.”

Craig’s message to the room was clear:

• Around 200,000 unrepresented taxpayers need support in the new system.
• Clients over £50,000 income should be signed up early, not in a last-minute rush.
• “Good bookkeeping has to be in place from the start of the year” if quarterly updates are going to work in practice.

In other words, MTD isn’t simply about policy. Bookkeepers play a crucial role in driving behavioural change.

For example, you can stop confusion around HMRC agent service accounts (the special logins agents use to manage client tax affairs), handling digital exemptions, and supporting clients who might otherwise delay action until problems escalate.

MTD exposes how dependent the system is on the human API layer you provide as a bookkeeper.

4. Bookkeepers already shape technology: Vendors follow you, not the other way around

Bookkeepers decide what tools survive, which workflows make sense, and ultimately, what gets adopted. You are the tester and validator, finding the first point of friction, or the first point of failure.

• If a tool creates more confusion or more admin, it dies.
• If a tool removes friction, it spreads fast—usually through bookkeeper networks, not via marketing campaigns.

You don’t adopt technology because it’s smart. You adopt it because it removes chaos.

Even on the Bookkeepers Summit exhibition floor, the pattern was clear: vendors chase bookkeeper workflows.

This is more evidence of the human API effect. The ecosystem moves to where bookkeepers need to remove friction.

5. Your role is expanding into leadership, strategy, and tech advisory

The bookkeeping profession is evolving faster than its training.

New pathways you’ll need to become familiar with are:

• Leadership of teams
• Change management
• Strategy
• Tech advisory and oversight.

This shows that the human API element is expanding further into a “human in the loop” intelligence layer.

Bookkeepers are no longer the last step in the chain. You’re becoming:

• Early signal detectors
• Workflow designers
• Compliance interpreters
• Trusted strategic guides for small businesses.

Bookkeepers are carrying more of the daily financial load than ever before—more clients, more real-time expectations, more admin, and more disconnected systems.

That’s why the conversation naturally turns to AI: not as a replacement for your judgment, but as the only realistic way to reduce rising operational strain without lowering your standards.

AI: Automation of chaos, not expertise

If bookkeepers are the human API, then AI has one purpose: to remove the friction created by systems that don’t talk to each other, not replace your expertise and oversight that holds everything together.

AI that promises anything else is a distraction.

Right now, you might spend too much time acting as a translator between mismatched systems.

You’re looking at HMRC portals, invoices, bank feeds, client behaviours, spreadsheet imports, late payments, mismatched data, and tools that were never designed to talk to each other.

As a human API, you carry the emotional and operational load of keeping clients’ financial lives running even when systems don’t connect.

The moment you give AI or automation even a small part of that load, the value of your judgment becomes clearer, not smaller.

Based on the Summit, three things stand out:

1. AI should give bookkeepers their 86 hours back

Many of the time-consuming issues that eat into the 86 hours chasing debt per year, and that send make 38 businesses a day go out of business, can now be automated with the right AI tools.

These include:

• Late payment recovery
• Invoice chasing
• Client clarification loops
• Email archaeology
• Manual checking and re-checking
• Reconciliation of errors caused by clients, not systems.

AI should automate the drudgery, not the judgment, which sits firmly with you.

AI can replace the broken middleware around bookkeeping: the duplicate imports, mismatched portals, missing documents, and manual corrections that shouldn’t exist in the first place.

If AI can compress those 86 hours into 8, that’s impact and value.

In an example of how effective this time saving could be, Accrual World implemented Sage Accounting and AutoEntry to automate data entry and document processing—saving a full day per week on manual admin.

This allowed the team to focus on higher-value advisory work and proactive client support, rather than repetitive admin.

Automation didn’t just save time. It enabled Accrual World to onboard new clients directly onto the cloud and plan workloads more effectively.

2. AI should stabilise data before it reaches humans

Bookkeepers shouldn’t be cleaning up:

• Badly scanned invoices
• Inconsistent dates
• Duplicate transactions
• Missing notes
• Miscategorised expenses
• Broken imports
• Half-complete data pushed through APIs.

You’re forced to fix these because the digital ecosystem is fragmented.

A good AI system becomes an early warning tool:

• Flagging anomalies
• Correcting obvious errors
• Structuring messy inputs
• Aligning data across tools.

It makes the human API layer cleaner, not heavier. For example: a system that auto-flags duplicate transactions before they clog your bank feed cleanup, so you spend less time firefighting errors and more time focusing on higher-value judgment work.

3. AI should enhance, not overwrite, human judgment

Bookkeepers are expanding into leadership, change management, and tech advisory.
Your value is highlighted in:

• Interpreting signals
• Advising small-business owners
• Spotting early risk
• Guiding adoption.

AI should surface patterns and predictions, but the interpretation—the part that affects real people—stays human.

You don’t need AI that does your job. You need AI that amplifies your role.

Evolution from human API to human-in-the-loop intelligence

If bookkeepers have become the human API for Britain’s small businesses, the next step isn’t automation. It’s more intelligence.

It’s not AI-focused, but a human-in-the-loop intelligence: a model where AI handles the friction and the bookkeeper handles the meaning.

Imagine opening an MTD quarterly update where 70% of the cleanup is already done.

Imagine opening a quarterly update where not only 70% of the cleanup is already done, but duplicate transactions are flagged, anomalies surfaced, missing descriptions inferred—and your job is simply to interpret what matters.

This is already happening in the sector.

ICB’s move into leadership, strategy, change management, and tech advisory is a recognition that bookkeepers now sit closer to the truth of the business than anyone else.

You see early signals: late payment stress, cash flow drift, pricing mistakes, supply chain pressure, and client behaviours that quietly erode resilience.

• AI can surface patterns
• AI can clean the data
• AI can reduce the 86 hours of wasted admin that cripple small businesses.

But only bookkeepers can interpret those signals and turn them into better decisions.

This is where the human API becomes a force multiplier.

When bookkeepers are supported, rather than sidestepped by AI, the entire small-business ecosystem becomes strong.

That’s how you get to the government’s ambition: if every small business grows by just 1% a year, Britain adds hundreds of billions to the economy by 2030.

The path to that isn’t policy.

It isn’t product.

It’s authentic intelligence. Human first, AI-assisted, API-powered.

5 things bookkeepers should do next: your human API checklist

If bookkeepers are becoming the human API for small businesses, your next step is to make that role intentional, rather than accidental.

Here’s a simple way to start.

1. List the tasks you only do because systems don’t talk to each other

Write down the jobs that exist purely because data is fragmented—for example:

• Manually importing bank statements
• Rekeying invoice data from PDFs
• Fixing duplicated transactions
• Explaining HMRC letters because clients don’t understand them.

This is your human API load. Seeing it on one page makes the real bottlenecks obvious.

2. Identify your “86 hours” of repetitive admin

Look at the last four weeks and highlight anything that repeats more than twice:

• Chasing clients for missing receipts
• Cleaning supplier descriptions
• Recoding miscategorised expenses.

This reveals where your human API role gets stretched. These are your first automation targets.

3. Upgrade one workflow with AI

It only takes automating one tasks to see the benefits. Choose a single task and test AI on it. You can try asking it to:

• Draft the first response to client questions.
• Summarise long email chains into action points.
• Generate invoice chaser messages in your tone.
• Clean transaction descriptions from bank feeds.

Don’t overhaul your practice. Prove improvement in one human API workflow first.

4. Define where you add judgment (your human-in-the-loop moment)

For the workflow you pick, mark the point where your expertise matters:

• Sense-checking anomalies
• Making a risk call
• Adding the explanation clients actually understand.

This is the judgment layer that AI can’t replace. The part of the human API that becomes more valuable as AI cleans the noise.

5. Communicate the value clearly

When using AI, don’t talk about speed. Talk about:

• Fewer errors
• Better visibility for clients
• Clearer answers
• More regular check-ins.

You’re not removing yourself from the process. You’re elevating your usefulness by stripping out the friction.

Final thoughts

Bookkeepers are the infrastructure the small business economy runs on.

You spot cash flow trouble before it becomes crisis. You turn policy noise into something clients can act on. You hold together systems that were never designed to talk to each other.

AI won’t replace that role. It will only increase your value.

Because once the friction disappears, the 86 hours compress, and the data arrives clean and connected, what remains is your unique value. You show your clients how to turn information into decisions that keep their businesses alive.