Predict how youruserswill respond

august is a human prediction platform powered by Digital Twins of your own users trained directly on your own data. Ask questions, run focus groups and forecast reactions without waiting.

Who Is This For

Modern product teams use august where they already work.

Each role brings august into the tool they already live in — Claude, Figma, GitHub, or a focus group — and shares what they learn with the rest of the team.

Product Manager

Claude
  1. 1Drafts a PRD for a new feature in Claude
  2. 2Gathers twin feedback via the august MCP
  3. 3Discovers gaps and wrong assumptions
  4. 4Ships a ready-to-implement PRD, faster

Designer

Figma
  1. 1Builds mocks for a new feature in Figma
  2. 2Gathers twin reactions via the Figma plugin
  3. 3Refines around predicted behavior
  4. 4Fewer redesigns after handoff

Engineer

GitHub
  1. 1Opens a pull request that changes a core flow
  2. 2The GitHub app runs it past twins on every push
  3. 3Sees which twins get stuck, and where
  4. 4Fewer regressions and bug reports

UX Researcher

august
  1. 1Starts with an idea for a product pain point
  2. 2Spins up an august focus group to test it
  3. 3Uses dissent to find a better answer
  4. 4Higher average NPS once it ships

Team Member

august
  1. 1Needs the CEO's honest take on an idea
  2. 2Asks the CEO's Digital Twin first
  3. 3Sharpens the idea over a few rounds
  4. 4Brings a stronger version to the real meeting
ChatGPTGitHubClaudeSlackFigma

One shared source of customer truth

Every role runs the same twins from the tool they already use, so a decision made in a PRD, a mock, or a pull request is one the whole team can see.

What you can do

The research that used to take weeks.

Stop guessing before you launch. Show a twin your pricing, your pitch, or a half-baked idea and find out what they actually think.

De-risk decisions

Test your idea on the real person before you launch it.

Every twin is built from how real relevant people actually behave with your product. Shoppers, app users, prospects, site visitors. Run your pricing, your pitch, or a new page past them and get answers in minutes, not weeks.

Would you pay $240 for these trail shoes?
Sawyer F.
Pressure-test concepts

Put one idea to a whole panel.

Ask a panel of twins the same thing at once. You understand what is working and what isn't side by side in minutes, without booking a single call.

We're building an AI inbox for founders. Would you use it?
Focus Group Report

Would you use an AI inbox?

Participants
12
Rounds
1
Generated
Just now
Executive Summary

9 of 12 said they'd be down to try it. The thing that got them was how robotic the writing was. A few liked how scannable it was for that reason, though. Their real problem is how much of a spam of emails they received in the first place, not how it gets sorted or written.

Collect insights instantly

Run a survey and get answers in minutes.

Ask hundreds of digital twins the same question and watch the answers come in as they respond. What used to take weeks of fielding takes a few minutes.

You surveyed 240 twinsWould you buy AI powered reading glasses if they were $350?
Yes47%
No33%
Maybe20%
How we validate

Built from real behavior. Validated against real outcomes.

Curious how we evaluate digital twins?The full method, the baselines we tested against, and where the twins still get it wrong.Read the blog post
How training works

Every twin is built in three layers.

  1. 1

    Interactions and behaviors are fed into our training pipeline to generate decision-making and personality weights for each individual.

  2. 2

    Each interaction generates memories, giving the twin perfect recall of what an individual has encountered.

  3. 3

    Reflections are built on top of memories, so each twin understands the individual's higher-level thinking and motivations.

Segments

Build any research panel from plain English.

Describe the people you want to understand and august gathers the right twins, then lets you interview, survey, or run a focus group with them.

Step 01 - Population

Start with every twin

All of your digital twins sit in one pool, each trained on how a real person behaves in your product.

Power users in analytics dashboard
Step 02 - Match

Type the group you want

Say who you want to focus on in plain English. august matches on how people actually behave, not demographics.

46 twins in this segment
Would a weekly digest replace your morning check?
32 of 46
said no
Step 03 - Ask

Ask them all at once

Interview, survey, or run a focus group with the whole segment, and every answer traces back to real sessions.

Capability comparison

Digital twins predict what personas and LLMs can't.

A focus group of hundreds of your customers, built from how real people actually behave on your site.

Feature
Twinsaugust
Personas
ClaudeClaude
Trained on a real individual
Captures behavior, personality, decision-making
Perfect recall of every action that person took
Memories and reflections retrieved at prompt-time
Accurate at predicting real customer actions
Full capabilityPartial, prompt-dependentNot supported
Pricing

Start with your vertical. Scale to your own customers.

Get started for free, then buy more credits only when you need them. No upfront cost. No subscription fees.

Full

Self-serve

Start with generic twins, then build twins from your own customer data.

$0upfront, only pay for usage

$25 for 500 usage credits


  • 100 free credits when you sign up. No credit card required
  • 500 more when you add a payment method, plus 1,000 when your first set of custom twins is ready
  • Custom twins trained on your customer data
  • Up to 500 twins
  • Buy credits in $25, $50, or $100 packs. Unused credits never expire
Get started

Enterprise

Account-managed pricing for larger deployments and custom data needs.

Custom

Custom credit rate and included credits


  • Everything in Full
  • Account-managed pricing and allowances
  • Larger volumes of twins
  • Custom data integrations
Contact us

Typical credit usage

Estimated usage per run. Actual usage may vary based on the length and complexity of the research.

1:1 conversation

~20 credits

Focus group

~250 credits

Survey

~800 to 1,200 credits

Credits are August's unit of usage. They are spent when August generates new work. Actual usage varies with the length and complexity of the research. Viewing existing chat transcripts or results does not use credits.

See how it works

FAQ

Questions product teams ask before trying august.

What is august?

August enables product teams to conduct user research with digital versions of their real users. These digital twins, as we call them, are built from real behavioral session data. You can run interviews, focus groups, and surveys with the twins built from people who actually use your product and get answers in seconds instead of weeks. Because each twin is built from what a person actually did, not what they say they'd do, it surfaces what your users really want.

How is this different from prompting a generic LLM to role-play a customer?

We beat generic LLMs in a head to head comparison when asking to predict behavior, see how we evaluate digital twins. If you describe your user to an LLM and tell it to pretend to be them, it tends to drift toward the average, agreeable answer. An august twin is tied to one real person's behavior and remembers what they actually did through memories and reflections, so it answers the way that person's history suggests. And unlike a generic model that's wired to agree, a twin will push back if your idea doesn't hold up.

Who is august for?

Anyone who has to make a call about their product and wants to know how users will react before they commit. That usually means product managers, designers, marketers, and founders: the people who would otherwise wait weeks to recruit and schedule users, talk to randoms or LLMs, or just ship on a hunch.

Is august part of FERMÀT or separate?

August is a standalone product. The underlying digital twins technology started inside the FERMÀT commerce platform. August opens it up beyond commerce to SaaS, mobile apps, and marketing sites.

How accurate are the results?

We beat generic LLMs in a head to head comparison when asking to predict behavior, see how we evaluate digital twins. We have multiple case studies in progress and hope to provide a more validated answer that tests against human counterpart responses. In the meantime, the best test is to try it with your own twins. Signup comes with free credits, so you can run a real question and see what comes back before you spend anything. You can also check out how we train digital twins.

What does august compete with?

Traditional user research: the Kantar, UserTesting, Nielsen, and Listen Labs approach of asking strangers or your coworkers what your customers think. If you're paying to survey strangers, you should be talking to twins of your actual customers instead.