An accounting firm.
The dozens of small steps behind a set of accounts: files merged into one, transactions pre-categorised, reports and filing documents prepared, data collected from clients by an agent instead of a phone call and a questionnaire.
Vision
First the data. Then the back office. Then the site. Then the robots.
Every business is two things: the back office and the delivery. Here's how we think both will change over the next ten years, in what order — and where we've chosen to stand.
Executive summary
A transformation is under way, and it happens in a set order. You can't skip a stage.
Stage one — centralisation. Before anything can be automated, the data and the systems have to be connected: one place where the numbers live, systems that talk to each other. Most companies haven't done this yet. It's the unglamorous stage, and it's mandatory.
Stage two — the back office, now until about 2030. Every routine operation in finance, admin, sales support, HR and reporting gets automated. The winners of this stage are businesses where the back office is most of the cost: estate agencies, accounting firms, agencies and services — anywhere people, not purchases, are the main expense.
Stage three — the "cyborg" stage, roughly 2030 to 2035. Once the back office runs itself, the people who actually deliver the work get more and more AI around them. Developers and data analysts already have it. Next come the trades: on construction sites we're testing augmented-reality glasses that show an electrician or an HVAC engineer the instructions, step by step. The idea is simple — let less expensive people do more complex work, with a certified expert supporting remotely, signing off the result and carrying the responsibility.
Stage four — robotics. After that, a lot changes — and we won't pretend to know exactly how. What we do know is that a robot can't work in a business that still runs on paper. Stages one to three are what make stage four possible.
The thread through all of it is decisions. Every stage is a chain of them — what to connect first, what to automate, which process to change, when to trust the machine. That's the part we've chosen to work on.
Stage 1 — Centralisation
What happens. All data flows and all system integrations are brought together: one source of truth for the numbers, systems connected to each other instead of through people re-typing things. ERP, CRM, finance, operations, marketing — joined.
Why it comes first. Automation on top of disconnected systems automates the mess. Every later stage — agents, AI on site, robots — stands on this one.
What we do here. Data platforms, integrations, one trusted source of data.
Stage 2 — Back-office automation (now → ~2030)
What happens. Routine back-office work stops being done by hand: data entry, reconciliation, reporting, approvals, document handling, scheduling, chasing. Agents do the repetitive steps; people handle the exceptions and make the decisions.
Who wins first. Businesses where the back office is the bulk of the cost and there is little procurement in the middle: estate agencies, accounting and bookkeeping firms, professional services, agencies, consultancies, recruitment. For them this stage is the whole game.
What it looks like.
An estate agency.
One manager used to look after 50–100 properties. With client communication, enquiry handling, notifications and document generation automated, the same manager handles 200–300.
What we do here. Custom operational layers on top of ERP and CRM, AI agents for each department.
Stage 3 — AI on the delivery side: the "cyborg" stage (~2030 → 2035)
What happens. The people who deliver the work — engineers, analysts, developers, and then the trades — get an increasingly detailed AI layer around them: instructions, checks, context, remote expertise.
On site. Construction is where this matters most, and where we're testing it now: AR glasses that show step-by-step instructions to electrical and mechanical trades, with a certified engineer supporting remotely and inspecting the result.
The economics. Less expensive people do more complex work. Certified people stop doing the work and start accepting it — remote support, sign-off, responsibility for the result. Skills scale; quality is controlled at the point of acceptance.
What we do here. Through Tellum, our construction company.
Stage 4 — Robotics
What happens. When the back office is automated and the delivery side is instrumented — instructions, measurements, acceptance all digital — the physical work itself starts to move to machines. A lot changes at that point, and we'd rather not pretend to know the details.
Where we are. Building the three stages that make it possible.
The thread: decisions
Each stage is a chain of decisions: what to connect first, what to automate, which process to change, which tool to trust, when to move. Companies that decide fast, small and recorded get through each stage faster — and arrive at the next one with the lessons written down.
That is why we are a Decision Intelligence company, not an automation vendor.
Our approachWhere we stand today
Stages 1 and 2 are our daily work.
Data Never Lies: centralisation, operational layers, AI agents, the decision process that holds it together.
Stage 3 is Tellum.
Our construction company — on site, now.
Stage 4 we are preparing for
by doing the first three properly.
Which stage is your business at?
Pick your first stepTell us where you are. We'll tell you what comes next.