AI Form Fill is powered by Storm, Sitemate's AI data capture agent, and works across Dashpivot Web, Dashpivot Mobile, and selected Sitemate Mobile workflows.
Storm's AI Form Fill feature automatically generates draft responses inside digital forms by processing spoken information, uploaded photos, and typed notes while maintaining existing review and approval workflows.
What AI Form Fill Is Used For
AI Form Fill supports field automation and reduces administrative data entry by helping teams with the following outcomes:
Drafting form responses instantly from spoken voice recordings.
Inspecting uploaded photos to automatically match and map them to the correct form sections or table rows.
Generating overviews and text summaries from unstructured notes or audio.
Populating structured data such as dropdown menus, linked lists, and logic-triggered fields without manual selection.
Uploading video footage — Storm extracts both the audio narration and visual content from a video to fill form fields, combining what's said and what's seen in a single pass. Available on Dashpivot Web only.
Filling People fields by matching names heard in audio or seen in a photo to real workspace users — Storm populates the field automatically when a confident match is found. If no clear match is found, the field is left for review rather than guessed. People field support is available on Dashpivot Web only; mobile is not affected.
Assisting with multilingual input by translating spoken data into the template's destination language.
Extracting data from field documents such as plant dockets, permits, and snag lists — photograph the document clearly and Storm will pull the relevant data into the form.
Attaching documents for context — click the paperclip button in the AI Form Fill canvas to attach PDF, DOCX, XLSX, CSV, or EML files. Storm extracts text from the document and uses it to fill relevant form fields. Storm can also read files already attached to the form as extra context — you don't need to re-upload them.
Referencing previously completed forms in the same folder as fill context — mention a past form in your directive (e.g., "use yesterday's site diary" or "same crew as the last pre-start") and Storm will draw on that form's content to help fill the current form. Storm cites which historical form informed each fill decision. Works across template types within the same folder. If the referenced form cannot be found, Storm tells you rather than guessing. Note: photo files from historical forms cannot be referenced.
Note: Users remain responsible for reviewing all generated content before saving or submitting forms.
Where AI Form Fill Is Available
AI Form Fill is available across Dashpivot Web, Dashpivot Mobile, and selected Sitemate Mobile workflows. Feature availability may vary depending on platform and workflow configuration.
How AI Form Fill Works: The Processing Stage
AI Form Fill uses Storm to examine a collection of inputs—including voice recordings, photos, typed notes, and existing form context like weather or location metadata.
When multiple inputs are provided together, Storm merges them into a single form entry, drawing the appropriate data from each source. For example, a voice note describing a defect and a nameplate photo will combine — the asset ID and manufacturer come from the photo via OCR, and the description comes from voice.
During the processing stage, Storm identifies activities, equipment, names, and issues, and maps them to supported fields. Storm operates on a confidence-based behaviour model — it will only fill fields when confidence is high enough. If Storm is unsure about an input or a dropdown match, it leaves the field blank for human review instead of guessing. Users always remain fully responsible for reviewing, signing, and submitting the final draft.
When you type a short, clear instruction — like "all pass" or "say yes to everything" — Storm acknowledges it and fills the form straight away, typically in under 10 seconds. Voice recordings, photos, and longer notes take a little more time as Storm works through everything in detail. You don't need to do anything differently — Storm handles this automatically in the background, and the experience looks the same either way.
See: How Does Storm Process Photos in AI Form Fill?
See: How to Use the Weather Field With Storm's AI Form Fill in Dashpivot
Typical AI Form Fill Workflows
Many teams begin with lower-risk workflows before expanding AI Form Fill into more complex operational processes.
Recommended Starting Workflows | Advanced Operational Workflows |
• Site diaries
• Daily reports
• Pre-start briefings | • Detailed inspections
• Defect reporting
• Site walkthroughs |
• Meeting minutes |
|
Supports Progressive, Iterative Workflows
Users can run AI Form Fill multiple times before saving a form, allowing documents to evolve progressively. For example, users can record additional updates later in the day, attach more photos, provide follow-up instructions, ask Storm to rewrite sections, or continue filling missing details.
When AI Form Fill is run on a form that already contains content, Storm reads the existing content and amends it based on your new input. Existing content is preserved unless you explicitly instruct Storm to remove or replace it.
How Storm Handles Fields With Existing Content
Storm's behaviour with existing content depends on what you've recorded or instructed it to do:
Fields that aren't mentioned or relevant to your input — Storm leaves these alone and does not change them.
Fields that are relevant to your input — Storm will amend or add to the existing content as appropriate, rather than overwriting it.
Fields you explicitly instruct Storm to overwrite or remove — Storm will do so only when given a direct instruction (for example, "replace the existing rows" or "remove this section").
This makes Storm well-suited for workflows where forms are cloned from a previous record — such as daily site diaries or shift reports — where some content carries over from day to day and only specific fields need updating.
Example — adding to an existing table:
A cloned form already has five rows in a labour tracking table. You tell Storm: "We also had Amy, Billy and Chad on site. They all started at 8 and finished at 6." Storm will add new rows for those people without touching the existing rows.
Example — replacing rows on explicit instruction:
You tell Storm: "Remove the existing rows in the table and replace them with new rows for Amy, Billy and Chad. They all started at 8 and finished at 6." Storm will remove the existing rows and add the new ones as instructed.
Example — amending a text field:
A multi-line text field already contains a bullet point list of activities. You tell Storm about additional activities completed. Storm will read the existing list and add the new activities to it, without removing what was already there.
To hard-prevent Storm from writing to specific fields entirely, field-level exclusions can be configured by your Org Controller. See: How to Manage AI Form Fill Permissions, Controls, and Settings
Governance, Permissions, and Safety Controls
AI Form Fill Does Not Automatically Submit Forms
AI Form Fill is designed to assist with drafting and completing forms, but users remain responsible for reviewing submitted content.
Storm does not:
Approve workflows
Sign forms
Complete signature fields
Automatically submit forms
(All required fields and approval workflows still apply).
Storm also never auto-fills signatures and sign-off fields, system-generated numbers (such as permit or document numbers), and prefilled template content. These always require manual input.
Permissions and Controls
Teams can control where AI Form Fill is used. Depending on the setup, AI Form Fill can be managed at the workspace level, template level, or field level. This allows teams to progressively introduce AI workflows while maintaining governance and oversight.
Current Limitations
AI Form Fill currently does not have the capability to:
Work inside Dashpivot Actions
Additional capabilities may be introduced in future Storm releases.
See: Which Fields Does Storm's AI Form Fill Not Support?