Honest and Grounded AI infrastructure for every research delivery team.
Honest and GroundedAI infrastructure for every research delivery team.
Metaforms helps market research agencies and sample providers standardize and scale delivery with agents that work alongside their team at every step, on workflows they already know and use.
Book a call
35+ research agencies use Metaforms for their research ops
Improve programming efficiency
Generate surveys in reviewable batches so programmers stay on the parts that need a decision.
Increase delivery capacity.
Take on more projects without adding headcount or stretching the team thin.
Protect quality and margins
Keep output reviewable and standards consistent across every delivery.
Streamline your delivery
Move work cleanly from intake to handoff, with fewer stalls between stages.
AI that works without disrupting how your team works.
Test project
Version 1.0.4
Upload survey.xml
Clarifications
3
Test Survey
Active Project: Brand Equity Tracker W1
survey.xml
routing_rules.json
8
9
10
11
12
13
14
<question id="Q5" type="single" required="true">
<title lang="en">Are you currently a student?</title>
<option value="1" label="Yes" nextQuestion="Q6" />
<option value="2" label="No" nextQuestion="Q7" />
<validation rule="required" message="Please select an option" />
</question>
[48 lines hidden]
63
+ <question id="Q12" type="single" required="true" category="demographics">
64
+ <title lang="en">What is your current employment status?</title>
65
+ <description>Select the option that best describes your situation.</description>
66
+ <option value="1" label="Employed full-time" score="4" />
67
+ <option value="2" label="Employed part-time" score="3" />
68
+ <option value="3" label="Self-employed" score="3" />
69
+ <option value="4" label="Unemployed" score="1" />
70
+ <option value="5" label="Student" score="2" nextQuestion="Q13" />
71
+ <option value="6" label="Retired" score="2" />
72
+ <validation rule="required" message="Please select your status" />
73
+ <skipLogic condition="value==5" target="Q13" />
74
+ </question>
75
+
76
+ <question id="Q13" type="multi" maxSelections="3" category="education">
77
+ <title lang="en">What is your highest level of education?</title>
78
+ <option value="1" label="High school diploma" />
79
+ <option value="2" label="Associate degree" />
80
+ <option value="3" label="Bachelor's degree" />
81
+ <option value="4" label="Master's degree" />
82
+ <option value="5" label="Doctorate (PhD)" />
83
+ </question>
84
85
86
87
<metadata version="2.4" lastModified="2026-08-05T09:30:00Z">
<author name="Research Team" email="research@example.org" />
<status>draft</status>
</metadata>
1 change pending review
Reject All
Accept All
Survey Programming Agent
Assigned to: Employment Demographics Block
AI
Metaforms Agent
11:32 AM
I have structured the employment status single-select block and mapped out all four standard demographic targets. Here is the generated XML output for validation:
survey.xml (Diff)
+ <row value="1">Employed full-time</row>
+ <row value="2">Employed part-time</row>
... (2 more rows hidden)
11:34 AM
You (Reviewer)
Looks solid. Let's merge these rows into main.
Planning next moves: setting up validation routing...
Looks solid. Let's merge these rows into main.
A useful tool fits the way people already work. Your team gets value from Metaforms without changing the platforms, templates, workflows, and review habits it already knows.
Our partnership has delivered strong ROI, thanks to Metaforms' exceptional service and prompt support. Their shipping speed is impressive, and we can always rely on them for timely deliveries, making them a trusted and reliable partner in our business operations.
Tabita Razaila
Head of Operations, STRAT7
Truth seeking is the job.
Everything else is just in the way.
The elephant in the room is:
bottlenecks are killing
your throughput.
The client deadline does not move. We need more time in the field, because that's where the challenge is, when we don't have time and the client has a deadline. I've got a project right now with seventy-five items and two thousand different attributes. It needs to go live tomorrow or the next day, and it's not possible to check everything that's in there. We get it out to QA and QA doesn't pick it up for a day. That's killing us. We're seventy-five percent through fielding when we discover the issue. We never have a standardized questionnaire. In one project, we received the questionnaire, an Excel file with the brand list, and a separate Excel file with the client's project workflow. Their Word docs are incredibly messy. They leave out instructions. When they do put instructions, it references the wrong question number. We have the doc that represents V1 of the survey, and then it goes through changes and revisions. There's bad maintenance of the document on our side. We hope the programmer used the naming convention according to the questionnaire. Otherwise it's messing up the syntax. A client will say, instead of simple least-fill, pick two brands on priority and put the rest on least-fill. They mention it in simple language, so a programmer, or anybody, can understand what's actually required. It always takes at least an hour to QA. We manually go through the link and check everything, then look at the data. They point out ten issues. We spend an hour chasing those down. They're not issues. If I have a wave and I want to make changes for the new wave, I do it manually. It's a large tracker study, with about a hundred and fifty menus, and they change every quarter. Someone has to manually look through those. If there are a hundred and fifty brands, that's fifty to sixty thousand lines of code. We've written a lot of trackers that change twenty percent from wave to wave. It's a little annoying. Our conjoints and MaxDiffs are never standard. We have something called a barrier MaxDiff, and something called a shift MaxDiff. Manually, that one would have taken around three years. None of our surveys are easy. We just had a single word translation mistake that cost eighty thousand dollars. The worst part of it isn't the money, it's the angry client. We have fewer programmers, but we're doing the same amount, if not more, programming. It feels like driving a Ferrari in rush hour. Everything in between gets squeezed. We have shifted the bottleneck. The deadline still hasn't moved.
Where teams start
Start with
Survey Programming.
For most teams, survey programming is the tightest bottleneck, and hence, the most logical place to start.
Our survey programming agent takes the questionnaire, applies your base template along with your custom coding conventions and rules, and generates code for 3-5 questions at a time.
You can accept or reject each proposed change, edit the output directly, or ask the agent to make the changes. The agent then carries these corrections into the next generation.
Work with agents on the survey platforms you already use.
Metaforms works with Decipher and Dimensions from its programming workspace. In Qualtrics and Confirmit, the agent works through an extension inside the platform.
MEDICAL PATIENt record
V1
MEDICAL PATIENt record
NEW V2
Confirm Q2 page placement? same page as Q1 and Q3?
Type your answer...
Q2 ROUTING
PENDING
0/4 addressed
Apply Changes
Manage trackers and revisions with ease.
Upload the revised questionnaire as the new source of truth. Metaforms compares it with the version already in programming and raises any clarifications.
It then shows proposed changes question by question, with controls that protect collected data and tracker history.
From the field
We've seen pilot users go from “this sucks!” to “it is useful for some things but still needs a lot of reviewing” to “it's helping us program surveys much faster” to “can you help us fix QA as well?”
Your team will need a few weeks of real use, templates, AI rules, and support before they see the full efficiency curve.
Where teams struggle
Then let the
next bottleneck
tell you where to expand.
Once programming moves faster, the next constraint becomes easier to see.
It might be questionnaire design, QA, tracker waves, validation, tabulations, or reporting.
Metaforms is built so you can add the next workflow only when it's clearly constraining delivery.
The Ferrari Problem
A customer told us that once his survey
programming got about 60% faster, his
QA stayed as slow as ever, so it felt like
driving a Ferrari through the middle of
Times Square. We think it's a lovely way to
describe how relaxing one bottleneck
reveals the next one.
Metaforms helps you adopt AI methodically, by fixing bottlenecks as they appear.
Roll out this way
and the gains compound
stage by stage as you expand.
You want your agents slow
and your team fast.
Metaforms is the most pragmatic way to scale research delivery with AI.
Book a call
Keep client revisions from sending your team back through the whole project.
For supported tracker workflows, Metaforms helps your team compare versions, review affected questions and logic, and protect IDs, labels, stored values, data layouts, quotas, and history.
Metaforms has been a game-changer for our survey programming process. By pairing our highly skilled programming team with Metaforms' AI solutions, we've improved efficiency while maintaining the quality and consistency we're known for. The platform has scaled smoothly across our 30+ person programming team, and with strong enablement and training from the Metaforms team, adoption felt practical and seamless.
Mark Scheckner
Senior Director, Market Research Data Operations, Prodege
Easy to work with.
Easy to trust. Easy to adopt.
Metaforms is the most pragmatic way to scale research delivery with AI.
Book a call
How teams work with us
Use Metaformsaround your current platforms,or let it carry the project end to end.
PROFESSIONAL SERVICES FLOW
Work alongside your current platform
Your surveys still live inside your current platform. Metaforms works alongside it, helping your team execute faster with reviewable output and shared standards across delivery work.
AGENTIC FLOW
Carry the project end to end
Start from a research objective or document. Metaforms helps turn it into research your team can review and launch, with human checkpoints along the way.
Receive real help from a real person, any hour, any day, within 30 minutes.
We take customer success seriously because getting AI right is mostly about getting people comfortable with new ways of working.
Metaforms support is available 24/7. A real person monitors support requests, and we aim to respond in under 30 minutes.
Slack or Microsoft Teams for account channels
A dedicated channel with our team, so getting help feels like pinging a colleague.
Support email and ticketing for issue traceability
Every issue logged and tracked to resolution, with status you can see anytime.
Help-center material and session recordings for rollout
Guides and recorded sessions your team can revisit as they roll out.
The Manual Nobody Reads
A customer told us most tools arrive with a 40-page manual and a goodbye. What got their team using Metaforms was someone answering in Slack within minutes. Turns out people adopt people, not software.
We've found Metaforms incredibly valuable for improving our programming efficiency. It's a great tool for our developers and programming team to have in their armour. We use it as a smart assistance tool, while always ensuring that our experts provide the direction and oversight. The customer service and support we've received from the Metaforms team have also been very professional and always on hand to help with any query our team might have.
Paolo Gambetti
Head of Innovation, Purdie Pascoe
Pricing
Pay by the project,
not by seats or tokens.
Add as many people to a project as you need and use Metaforms agents as much as the work calls for. Pricing follows the project. Seats and token usage do not drive the bill.
Book a call
On metered pricing
One customer said metered pricing “feels like a penalty if we use it more than we think we will.” We think that's exactly right. So we price by the project and let you actually use the platform. No one on your team should have to think twice about whether running one more check, adding one more teammate or generating one more draft is going to cost extra.
Security
Your security team won’t be our first.
Metaforms has cleared infosec reviews at some of the world’s largest research agencies and panel providers. Our security posture is independently verified through ISO/IEC 27001 certification and a SOC 2 Type II report. We also commission an independent penetration test every year.
SOC 2 Type II Compliant
ISO/IEC 27001 Certified
Annual penetration test
Enterprise infosec cleared
Your data never trains or improves any AI model
Every AI output requires user review and approval before use.
Your data stays isolated
Application and database controls prevent access across customers.
AES-256 at rest. TLS 1.2+ in transit
Every endpoint is HTTPS-only.
Enterprise SAML SSO, with MFA support
Role-based permissions limit each user to approved resources.
Automatic failover across multiple availability zones
Continuous backups support point-in-time recovery.
Sub-5-minute incident response
Around-the-clock monitoring targets detection of critical issues in under 5 minutes.
Every code change is reviewed and tested
Dependencies and software supply chains are scanned before deployment.
US or EU data residency support
Set retention periods and permanently delete data when required.
Proud to be part of the launch of Konovo's Survey Studio product and the development of our partnership with Metaforms. Survey Studio will be immensely useful for our customers to improve and accelerate the process of gathering healthcare insights, while laying the groundwork for further AI-assisted workflow improvements.
Steven Blum
Global Corporate Development Principal
Pilot
Test Metaforms on work your team knows well.
Start with a real questionnaire your team already understands. We can run a side-by-side benchmark, share the output, timing, and review notes, then shape the pilot around what your team needs to prove.
