INSIDE SITEMATE

Zero Click Forms

Preface

For anyone who’s ever spent time on a site within the built world, or in the back office that supports one - they know the truth;

The work is physical, but the proof is in the paperwork.

For a long time, that split was manageable - you’d do the work, you’d get paid, simple.

Then, things started to become complicated.

The history of information management within the built world can be broken down into three eras;

  • The faxing era; when engineering drawings were faxed and posted between parties.
  • The file storage era; when companies like Aconex, QA Software (TeamBinder) and Procore were founded - were verticalized/industry-specific file storage platforms were created.
  • The database era; where the built world started to realize that the value was not so much in the file storage, but from within the data within those files.

The covid-19 pandemic was the catalyst for monumental change within the built world, perfectly timed on top of multiple converging trends, all of which were driving the volume of paperwork, forms and files up, up and up;

  • Material and labour costs rising; The cost of building ‘real things’ is rising across materials, labor, and logistics, shrinking margins and increasing scrutiny on every decision. As inputs become more expensive, tolerance for rework drops to near zero, driving heavier documentation around quality, approvals, traceability, and cost justification to protect margin. This drives more documentation around approvals, quality checks, traceability, and cost justification to prevent rework and protect financial outcomes.
  • HSEQ regulation; Health, safety, environmental, and quality regulations continue to expand in scope and enforcement. Expanding safety, environmental, and quality regulations directly translate into more mandated forms, records, permits, and evidence - documentation is no longer optional, it’s the mechanism of compliance.
  • Number of parties involved; Modern projects involve an increasing number of subcontractors, consultants, suppliers, and authorities. Every additional interface creates new coordination points, handovers, approvals, and contractual obligations - each one generating its own layer of reporting, sign-off, approvals, and record-keeping.
  • Brownfield projects, sites and work; Brownfield projects involve building around, modifying, or integrating with existing live infrastructure. The risk profile is significantly higher, which results in far more planning documents, risk assessments, permits, isolation procedures, and change records than greenfield work.
  • Gen Z entering the workforce; As a new generation enters the built world, average experience levels are dropping and workforce turnover is increasing. To manage this, companies lean heavily on documentation for onboarding, training, competency sign-offs, and audit trails to replace what used to live in people’s heads.

Below is an excerpt from Sitemate’s investor pitch deck in 2024, that summarizes this trend.

These forces compound, not linearly but exponentially, and documentation becomes the pressure valve that the industry uses to manage risk.

Covid was the inflection point that exposed how fragile these paper-heavy, people-dependent systems really were; when teams were suddenly remote, sites were fragmented, and physical handovers broke down overnight. Digitization wasn’t a transformation initiative anymore; it became the only way to keep work moving, risks managed, and accountability intact in a world that could no longer rely on being in the same place at the same time.

Here’s the part I joke about, but it’s also painfully real:

A lot of enterprise buyers don’t have an intrinsic sense of urgency about technology decisions because they’re senior leaders who came up in a time when there simply wasn’t this much paperwork.

They remember a world where projects ran on radios, clipboards, and tribal knowledge. The documentation burden existed… but it didn’t swallow their working day.

They haven’t felt the problem firsthand.

But the people on the tools right now feel it. The supervisors feel it. The engineers feel it. The administrators feel it - and the curve keeps on compounding, the pressure keeps on compounding.

A Necessary Evil

Whether one likes it or not, the built world is all about records, proof, documentation.

We’ve managed to consistently improve our NPS over the years, but every time we reach a ceiling, there is always part of me that wonders - can we even get it higher, when paperwork and forms is something that no-one wants to do in the first place, and in some cases, where they either hate doing it, or physically can’t due to language/literacy issues?

And then somehow, eventually, we’ve always found a way to raise it higher.

  • Records in the built world are;
  • Proof something happened (or didn’t).
  • Evidence that standards were followed.
  • Traceability when something goes wrong.
  • Accountability when money changes hands.
  • Continuity when teams rotate and projects stretch for years.

Forms are not optional, they are the operating system of modern construction, infrastructure, maintenance, and heavy industry.

But they come with a brutal paradox:

The more society demands safety, quality, and accountability,

the more paperwork gets pushed onto the people closest to the work,

the people with the least time to do it.

Sitemate’s Mission

At some point in a product’s lifecycle, you stop needing a roadmap and start needing a compass.

Because when you get early success, and more resources, you can then build anything, and that in itself is the problem - building the right things, and knowing why you’re doing it.

Furthermore, AI has made the number of options available explode.

In order to navigate this maze, Sitemate needed a compass - I needed it more than anyone.

So I re-wrote our mission until it was clear enough to become our compass.

If you’re reading this, then you should know by now that Sitemate’s Mission is to Multiply The Engineering Power Of The Built World;  Multiplying The Engineering Power Of The Built World — The Sitemate Manifesto.

This means that we aren’t just going to ‘take people digital’, or streamline things, we are going to unshackle our engineering workforce and free them of their burnout and their burden, so that they can overcome the built world’s largest challenges.

Our Manifesto details the personal implications of this load, on people within the built world, and there’s further background here on how it was written and formulated here;  Under The Hood — Of The Sitemate Manifesto and Mission.

At the 2025 Offsite Kick Off in Phuket, I introduced Sitemate’s mission, and then followed up with how this will apply to our PDE (Product, Design, Engineering) teams - below is the slide;

Above you can see what I laid out to the team - in order to make progress against our mission, in order to combat the compounding curve shown above in this blog, we need to both;

  1. Drive usage up; in other words, users are using our products for an increasing number of use cases and instances of executing a given use case.
  2. Drive the time taken to execute a task down; which could be measured in either Clicks to Complete (CtC) or Time to Complete (TtC), for example - a photo batch or a form completion.

… at the same time.

Our AI Journey

With the arrival of the AI Era, like it should have done for every CEO - it became apparent to me that Sitemate needed to evolve, which in itself was a rather strange realization for me personally.

I remember thinking;

“Wait, aren’t we a startup?

Why do we need to evolve?”

The short answer was;

  • No, we’re not a startup anymore - Sitemate was founded in the SaaS era, and was, at least at the time of this realization, by definition, not AI-native. Hence, a re-founding journey and event is required.
  • Yes, we need to evolve - and if we don’t, Sitemate, like every other technology company, is at risk of falling into irrelevance.

That being said, in these situations, it can be very easy to simply get attracted to shiny objects, to just tick boxes and appear to be ‘keeping with the times’, i.e. in this case; it equates to ‘AI washing’ your product and brand.

At the start, the only thing that Sam and I were 100% certain of, is that we didn’t want to just add a shiny AI button, or sprinkle some pixie dust across the product.

But, forming a clear vision and building conviction in Sitemate’s path ahead was certainly not a straight line.

During the 2024 Founders Offsite in Cape Cod, we landed on our first strong hypothesis for Sitemate’s path forward.

The vision was centred around an agent that could do the work of our users, do specific tasks for them instead of them doing it themselves.

We listed out every possible manual task that we could think of, and tried to cluster them under Jobs To Be Done, shown below.

From there, we simply picked a task to start with - and for that, we landed on Material/Labor data importing from dockets (AU/NZ) / notes (EU) / field tickets (US/CA), which is essentially the manual process of reading a paper slip, often from a carbon copy triplicate book, and then typing it into your form.

An example of a series of dockets/notes/field tickets is shown here;

In order to have AI execute these tasks for you, Sam and I believed that the agent needed to hold some human properties, e.g. a name, a face.

We worked through a number of concepts, many of which were quite comical.

From there, we got to work - we started building; “Agent Sam”.

And, well… it worked!

Below is a screenshot from the early Alpha usage - users could open a Dashpivot form, click a button to engage Sam, give Sam a batch of concrete dockets, and Sam would extract the docket data and complete the table inside of a Dashpivot form for them.

From there, the learnings from the market came fast;

  • We learned that the breadth of use cases was significantly more distributed than our original hypothesis - users didn’t want to just use this for concrete dockets, they wanted to use it to extract data from name plates on assets, as well as labels on manufactured products.
  • We learned that trying to integrate the agent into our UI directly, and replicate human behaviour, such as typing on a keyboard, was extremely time consuming and created fragility.
  • We learned that needing extensive configuration in the template editor, and mapping columns to AI agents was going to add significant complexity to the setup and configuration process, and we weren’t convinced this was going to scale well.
  • It was clear that there was merit in this path, but our v1 wasn’t optimal.
  • Our v2 vision attempted to solve these problems by introducing flexibility and the concept of an ‘agent builder’ - our running assumption was that the breadth in use cases could be solved by switching to a no-code style agent builder, and moving this into Flowsite.

However, after some initial scoping and exploratory work - it also became clear that this wasn’t the path; it would have been a huge amount of effort to simply build the pre-requisite capability to support agent building, such as UI/button control inside of Dashpivot, from within Flowsite.

Something didn’t feel right, and our Microsoft integrations were picking up serious momentum, so I made the decision to pause all work on the v2, take a step back, and reset.

We had kept our AI team relatively small during this phase, as there was so much uncertainty - we didn’t want to distract the core, and we didn’t want to over-invest, before our direction was clear, despite feeling like we were losing time.

Waging A UX War, Against Ourselves

With the learnings from our early efforts on our AI journey, and now guided by our Mission as a compass, I revised our framework, and landed on something that started to click.

At the July Allhands in 2025, the below framework was introduced to the team - we needed to attack our own core user experience, we needed to completely re-think what it meant to use Sitemate’s products, in the AI era.

This was when the mental unlock really started to kick in - we put our own UX on trial, not because it was bad, but because it was trapped in the old assumption.

Forms UX still expected “clicking and typing”.

And once I saw that clearly, I couldn’t un-see it.

What if forms didn’t need to be filled?

What if they could be produced as a side effect of work happening?

From here, the thinking started to spread like a wildfire amongst the most senior members of the Sitemate team - below are a couple of extracts from my Slack DMs with Eric, our Chief of Staff, as the spool of thought started to unravel in real-time.

Conviction built quickly, and this was followed by formalizing the AI Platform team, as well as bolstering the hiring roadmap, so that we could start to move into full scale execution.

Going Multi-Modal

For most of computing history, “filling a form” meant;

  • Keyboard
  • Mouse
  • Structured fields
  • A human “translating reality” into boxes

That made sense when typing was the only reliable input method, but then some things changed;

  • Voice became practical; We went from “voice-to-text is a gimmick” to “voice is a normal interface” in a shockingly short time. But there’s an even bigger shift than voice-to-text.
  • Conversation; Phone calls → conversation → transcript. Now: conversation with an agent → transcript → structured output. This matters because conversation is how humans naturally describe work.
  • Photos and videos became read-able by computers; On site, photos are often more truthful than text. A quick walkthrough video can contain more real signal than a 40-field checklist. The old workflow treats media as an attachment. The new workflow treats media as the input.

Multi-modal models changed what “input” means, and once an AI can combine;

  • Speech
  • Text
  • Images
  • Video
  • Files

With context for a given user (project, site asset, location, spec, historical and recent records), then it becomes conceivable to generate a form without a person filling it.

Target UX Flows

Below is a collection of the workflows we are targeting for our users, in-line with my zero click forms vision.

Daily Reporting

In simple terms - the vision is for a user to finish their day on site, get in their truck, get in their ute, and have a phone call style conversation about the day, maybe continuing with a draft of things that they’ve already completed earlier in the shift, such as photos or notes.

The agent would guide them through the questions or blanks, use it’s prompts to dig into certain topics and issues, and refer back to previous days' records, e.g. “Oh I see you’re mentioning a delay, is that the same delay as the one you reported yesterday, or a new one?”.

Input:

  • 60–120 seconds of voice: “Here’s what happened today…”
  • Or, a phone call back and forth
  • A batch of photos (progress, deliveries, key areas)
  • Optional: quick prompts from the agent (“Any incidents?” “Any delays?”)

Output:

  • Completed daily report template (structured)
  • Highlights + risks
  • Tagged subcontractors / work areas
  • Weather pulled automatically
  • Material deliveries cross-checked against dockets

Why it wins:

  • It allows for daily reporting to be completed during previous ‘dead time’, all whilst creating much higher quality records.
  • Diaries are high-frequency, low-love, and typically always late.

Inspections

The vision here is for a user to be able to talk through their inspection as they’re going, adding a single voice monologue and photos progressively as they go, rather than stopping and looking down continuously at their phone.

Input:

  • Photo/video capture of the inspected element
  • Agent asks targeted questions based on the checklist + spec:
  • “Is the batch number visible?” “What’s the measured tolerance?”
  • If measurement required: user speaks the value (or it’s read from a gauge)

Output:

  • Inspection form completed
  • Non-conformances flagged
  • Evidence attached and indexed
  • Corrective actions generated as tasks

Why it wins:

  • Quality is evidence-heavy and consistency matters more than verbosity.

Time Capture:

The vision here is to allow for users' time entries to be followed up automatically by an agent, and/or collected over the phone, using PIN verification.

At present, admins with large labor workforces often spend long periods of time chasing up people for times - imagine if an agent could simply monitor, or be triggered from an overdue action item, to simply call all of the workers via the Sitemate App to collect their time data.

Input:

  • Passive signals: location, schedule, crew roster
  • Lightweight confirmation (“You were onsite 6:48–15:12; confirm?”)
  • Voice exceptions: “Left early due to…”

Output:

  • Timesheet / labour allocation
  • Cost codes suggested automatically
  • Discrepancy alerts without admin chase-ups

Why it wins:

  • Time capture is pure admin drag and error-prone at present.

Inductions:

Workers cannot start work on site until they’re inducted, and at present, it’s a frictional process - even if the webform submission, and ‘remembering your details from last time’ problems are solved, there’s still major blockers.

For example, a large proportion of the workforce is illiterate in English, and many in their own native language (e.g. Spanish).

Input:

  • Human speaks their details or replies to questions
  • Agent runs a short conversational induction
  • Captures acknowledgement with minimal friction
  • Language selection / translation automatically

Output:

  • Completed induction record
  • Competencies and tickets extracted
  • Gaps flagged (expired certs, missing items)

Why it wins:

  • Solves for both accessibility and speed, at the same time.
  • Allows safety managers to reduce the rate of on-site inductions, which slows down the time before someone can start work.

Morning Briefings:

The vision here is to allow the foreman running the morning briefing to hit record at the start, talk through everything, scan everyone on, and then upon completion, have the audio file / transcript convert into populated form fields.

Input:

  • Supervisor voice briefing (recorded once)
  • Crew confirms understanding with 1–2 interactions
  • Optional: photo of whiteboard / plan

Output:

  • Brief record created
  • Attendance captured
  • Hazards and controls logged
  • Action items assigned automatically

Why it wins:

  • Pre-starts are repetitive, but critical when things go wrong.

All of the above being said, Zero click forms will only work if;

  • The agent is context-aware (project, asset, role, history)
  • The system can ask good questions (not generic prompts)
  • Outputs are auditable (who said what, when, based on what evidence)
  • There’s a clear correction loop (humans can fix fast, and the system learns)
  • It fits the real environment (gloves, noise, sun glare, time pressure)

Closing - The Built World Double Jump

Every so often, an industry doesn’t modernize in neat stages - it leapfrogs.

A classic example: many African markets didn’t go “landline → PC → internet → smartphone”. They went straight from landline to smartphone, skipping PCs, because mobile wasn’t a luxury upgrade, it was the first interface that actually worked with the constraints on the ground: cost, infrastructure, distribution, and accessibility.

I believe the built world is sitting on a similar edge, of making a double jump.

The conventional story says; first we digitize paper forms, then we perfect digital workflows, then we automate. But, this assumes the “digital forms era” is the natural destination.

Arguably for many workers and use cases, it isn’t.

Digital forms still assume the human should do the hardest part: translating reality into structured fields. That might work in an office, and for qualified engineers in the field, but it’s really a terrible mismatch for a jobsite, a shutdown, a plant, a fleet yard, or a tunnel project at 2am.

AI flips the sequence.

The built world doesn’t need ten years of “better form builders” to reach the future. It can skip the middle chapter and jump straight to a future where documentation is captured as a byproduct of work; a voice brief, a walkthrough video, a few photos, and a system that knows the context well enough to ask the right questions and produce an auditable record.

That’s the double jump - a large portion of the built world will go from physical paperwork to zero click forms, skipping the digital form filling step entirely.

Sitemate is perfectly positioned to be the company that unlocks this.

About the author

Hartley Pike

CEO & Co-founder

LinkedIn

Education:

University of Technology Sydney - Master of Business Administration in Entrepreneurship (Dropped Out)

University of Technology Sydney - Bachelor of Engineering (BE), Diploma in Engineering Practice, Civil Structures

Bio:

Hartley is the Founder and CEO of Sitemate, where he turns real-world construction and infrastructure experience into software products that site teams want to use. With a background in civil engineering across major transport infrastructure projects, manufacturing, and large-scale site operations, he brings experience to product strategy and development.

He applies lean principles to improve productivity and efficiency across complex, high-risk operational workflows. Today, he leads Sitemate’s product strategy and growth, building platforms such as Dashpivot, Flowsite, Gearbelt and more that transform standard operating procedures into structured, data-driven workflows, helping teams deliver work more reliably, safely, and on time.

Past Projects

Version

Last published

25-11-v2

07/01/2026

Want to learn more about Sitemate?

Visit about page

Visit careers page