AI is already here.

AI is a powerful tool. Like every powerful tool in history, it is only as good as the people using it. Humans are better. Humans will always be better. What changes is how much time they spend on things that are beneath them.

The moment we are in

Something shifted in the last three years that most businesses outside Silicon Valley have not fully registered yet. The ability to build software that understands language, processes documents, handles phone calls, queries databases and makes reasonable decisions without explicit programming, that capability went from research paper to production tool in a window most people blinked and missed.

We are not talking about ChatGPT in a browser tab. We are talking about systems that run inside your business, connected to your data, operating on your workflows, twenty four hours a day, at a cost that would have been unthinkable five years ago.

The founders and operators who have started using these tools properly are not talking about it at conferences. They are quietly getting three hours back every day. They are processing in a week what used to take a month.

Most businesses are not those companies yet. Not because they are behind, but because the people offering to help them have mostly been selling a vision of AI that puts the machine at the centre and the human at the margins. Dashboard demos. Proof of concept engagements that never ship. Chatbots that treat a customer like a query to be resolved.

The gap between what AI could do for people and what most businesses have experienced is enormous, and it is not a technology gap. It is a values gap. That gap is what sizrok was built to close.

The question is not whether AI will change how your business runs. It already is. The question is whether the version that touches your business was built for you.

What is wrong

The culture around AI right now is being set by a small group of people who are extraordinarily powerful, largely unaccountable, and building for a future that serves them first. We think that matters. Not as a political statement. As a practical one.

The big labs are moving at a pace that outstrips any meaningful accountability. Organisations with the technical ability to reshape how information is created, how decisions get made, how labour works, are operating with almost no external oversight. The people running them have enormous resources and a strong conviction that they are the right ones to determine what comes next.

We are not anti technology. We use the models these organisations build, and some of them are genuinely excellent. But we believe the culture around them, the technofeudalism, the accelerationism, the messianic tone, the paternalistic safety framing that still concentrates power in a handful of labs, is not the only way to build with AI. It is just the loudest one.

The investment culture compounds it. Billions of dollars have been deployed into AI companies that are building for exits rather than for people. The result is an ecosystem full of products optimised for demo performance and investor narratives, not for whether they actually make working lives any better.

And at the bottom of this stack sits the average business owner, being pitched by agencies selling "AI transformation" at rates that suggest the transformation in question is mostly of the agency's bank account. Slide decks. Workshops. Pilot projects. Retainers that keep the consultants comfortable and leave the client with nothing that runs.

This is the culture we are working against. Not because we want to make a political statement. Because it is bad for the businesses we want to work with, and because we think there is a better way to build.

Technology concentrated in a few hands has never ended well for everyone else. We build AI that is owned by the people using it, understood by the people running it, and accountable to the people it affects.

Where this goes from here

We are honest about uncertainty. Nobody knows exactly how the next decade unfolds. But we have a clear view of what the next few years look like for most businesses, and we are building sizrok around that view.

Now / 2026

The early adopter advantage opens up

A small percentage of businesses are running AI augmented workflows with a measurable speed and cost advantage. Most of their competitors do not realise this is happening yet. The gap is real and widening. Building working AI systems now is not just saving time today. It is institutional knowledge that will compound for years.

1 to 2 years / 2026 → 2027

Operational AI becomes table stakes

The businesses that have been building quietly will start to look dramatically different from the ones that have been evaluating. Response times, processing speed, capacity per headcount, these metrics will diverge sharply. This is also when the first wave of AI native competitors puts meaningful pressure on incumbents in services, logistics, customer ops and finance.

3 to 5 years / 2028 → 2030

Most processes have an AI layer

Not because AI replaced people. The nature of the work changed. Data entry, document processing, first pass analysis, inbound triage, routine reporting, handled by systems. The human work becomes the judgement, the relationships, the escalations, the strategy. Teams who navigated it well will be doing more meaningful work; the rest will be playing catch up.

5 to 10 years / 2030 → 2035

The question is who controls the infrastructure

This is when the ethical dimension becomes a practical business concern. Businesses on proprietary platforms, with data locked into vendor ecosystems, will find themselves dependent on the commercial decisions of a few technology companies. The ones who own their systems, understand their data, and can swap components as better options emerge will have a structural resilience the others lack. This is why we build for ownership from day one.

We could be wrong about the timeline. These projections are based on what we see in the businesses we work with and the direction of the underlying technology. Some sectors will move faster than others. But the direction is not in question.

Our view is that the best time to build your first AI system is now, while the competitive advantage is largest and the cost of getting it wrong is still manageable.

What we build

We are an automation studio. That is the core. The data layer, the voice systems, the internal tooling, they all exist in service of one idea: AI that takes the repetitive work off your team's plate and gives them their time back.

Workflow agents — daily tasks that drain your team

Invoice processing, lead enrichment, status reports, document classification, meeting notes, weekly summaries, exception logging. Not trivial problems. They compound across a team of ten into hundreds of hours a month, all below the level of the people doing them. We build agents that handle this in the background, with human review where it matters, and hand the time back.

Data layer — data that answers questions

Most businesses are sitting on more data than they can use. CRM records, sales history, support tickets, financial reports, operational logs, the answers are there, buried in spreadsheets and dashboards only one person knows how to use. We build a layer on top of your existing data that lets anyone ask a question in plain English and get the answer, the chart, or the report back.

Voice and inbound — agents that handle the phone and inbox

Your best person has a finite number of good hours in a day. An inbound AI agent does not. It qualifies leads at two in the morning, answers the same question for the hundredth time without frustration, books meetings, gathers intake, and escalates only the conversations that need a human. Trained on your tone. Connected to your CRM. Hands off cleanly.

Internal tooling — software your team will actually use

There is always a thing your team Slack asks each other ten times a day. A report someone has to pull manually. A process that lives in one person's head. We build these as custom internal tools, lightweight, fast, connected to your data, designed for exactly your workflow. No SaaS subscription, no features you do not need, no data on someone else's server.

How we build it

We have one process. It works on every engagement we have taken. It skips the parts that do not matter and gets to something real, running in your business, as fast as we know how.

One real conversation first. Not a scoping call where we gather requirements for a proposal. An actual diagnostic where we map your day together, find the three or four places AI genuinely helps, and identify the one with the best ratio of pain to buildability. You leave with a written recommendation whether or not we work together.

Fixed price, fixed scope. We agree what we are building, what it costs, and when it will be done before we start. No time and materials. No hourly rates that spiral. No scope creep negotiations. If we discover something harder than expected, we absorb it. That is what fixed price means.

Embedded working style. We work like a member of your team. Daily updates where they help. Weekly demos on working software. Direct Slack access. No agency intermediaries. The person on the call is the person writing the code.

You own everything. Code, prompts, documentation, the playbook for running and modifying the system. There is no version of our work that requires ongoing access to sizrok to function. You can run it, change it, hand it to a developer, or throw it away, without asking us.

We stay if you want us to. From the second month onwards we offer an optional, rolling, cancellable retainer for teams who want continuous improvement. Optional. Rolling. Thirty days notice. Never required.

We do not replace your people. The systems we build take on the tasks that are beneath your team, not the team itself. We automate the Monday morning inbox sort, not the account manager reading it. The human work stays human. The machine work gets machines.

Our principles

These are not values on a wall. They are the actual decisions we make when they cost us something. A principle that is only easy to hold when nothing is at stake is not a principle.

1. AI augments, never replaces

We decline work where the clear intent is to use AI to reduce headcount rather than to improve what the existing team can do. The systems we build make people more capable. They do not make people redundant.

2. Builders are accountable

We document every system we build. We explain how it works, what it decides, and where it can go wrong. No black boxes. If a client cannot understand, at a reasonable level, what their AI is doing and why, we have not finished the job.

3. Power should not concentrate

We build for independent ownership. Every system is structured so the client can run it, modify it, and migrate away from any underlying provider without rebuilding from scratch. No proprietary lock in. No dependencies on sizrok being there.

4. Transparent, always

We explain our technical decisions in plain English. We surface tradeoffs before we make them, not after. We do not add complexity to justify our fees. Could we explain this clearly to the person paying for it, and would they agree it was the right call?

5. Work we can do well

We limit the engagements we run at any time. Not for manufactured scarcity. Embedded work at the standard we want to deliver takes real attention. If we cannot give a project what it needs, we say so and point you elsewhere.

6. No data beyond the task

The systems we build collect what they need to do their job and nothing more. We do not aggregate client data across engagements. We do not train models on client data without explicit permission. A trust position, not just a privacy one.

Who this is for

For the operators and founders who have a real business, a real problem, and are tired of being sold theatre when they came looking for something that works.

You do not need to be a technology company. You do not need a data team or an engineering function. You need a problem worth solving, a process that eats time, a question your data cannot answer fast enough, a communication volume your team cannot sustain manually, and the willingness to actually ship something rather than evaluate it indefinitely.

We work with founders running ten person companies and operators managing hundred person teams. Service businesses, professional firms, consumer companies, B2B operations. What our clients have in common is not sector or size. It is that they have a specific pain point and they would rather fix it than talk about fixing it.

We are not the right fit if you are looking for a strategy engagement, a market analysis, or an AI readiness report. We are builders. We are a good fit if you have identified a specific operational problem and want someone to build a system that solves it, hand it over completely, and optionally stick around to improve it.

The first conversation is free, genuinely diagnostic, and ends with a written recommendation regardless of whether you want to work with us.

We take on a small number of engagements at a time. If the timing works, we would like to work with you.