Understand your people.
Make change work

Agent-led interviews uncover what people need, what holds them back, and where change can begin.

What's missing between rollout and routine?

First pilot roundFreeCompanies with up to 250 employeesQ3 to Q4 2026

How Fathomly works

Listen closely.
See the bigger picture

A useful insight starts in someone’s working day. Follow their words from an adaptive interview to shared patterns and a next step.

01 / Interview

Start with the work.

A follow-up reveals why the same project update ends up in several places.

Inside an interview

Fathomly asks

How do you usually keep people updated on your projects?

01Employee · Interview 01

I update our project board. Then I email my manager and post an update in Slack.

Fathomly follows up

What goes into the email and Slack message?

The same update, mostly. My manager asks by email, and the team checks Slack.

An excerpt from a 5–20 minute conversation.

From individual words to shared understanding. Fathomly connects the answers to project reporting, duplicate reporting, and reporting expectations. Six interviews reveal repeated updates, information needs, and a contrasting experience: one team uses the board directly in meetings. A possible test is one shared update with notifications in existing channels. Then ask whether anyone had to prepare the same update elsewhere and whether managers had the information they needed. Group reporting requires at least five participants.

Guided pilot

Is this software worth rolling out?

Before you commit to expensive software.

In a guided pilot, we scope interviews with you during the software trial and review the findings together, independently of the software vendor.

Discuss a trial pilot Group findings are released for five or more participants.

New project tool. Same spreadsheet.

Fathomly, interview agent: When did you last use the new project tool?
Employee: I updated it before the team meeting, then the spreadsheet too.
Fathomly, interview agent: Why both?
Employee: The other teams use the spreadsheet. The new tool adds a step.

Before buying company-wide:

Resolve the duplicate workflow, then test whether the software improves the work.

Employee experience informs whether to buy, adapt or not roll out. Your company decides.

What you can learn

What’s helping. What’s holding you back

These examples focus on AI adoption today. Extending this approach to other tools and ways of working is part of our direction.

AI opportunities

What would our teams automate first?
Is sales still writing proposals by hand?

Tool needs

Which team needs specialist AI, not ChatGPT?
What does Legal use when ChatGPT is off-limits?
Can HR use AI with personnel data in play?

Everyday AI use

Are our developers using the coding AI?
Does AI copy still sound like our brand?
How far along are we really with AI?

Shadow AI and safety

Is customer data ending up in public AI chats?

New tool rollouts

What's missing between rollout and routine?
Do our people see AI as help or threat?
Which teams are moving with AI and why?

Training and tool value

Which trainings do our teams really need?
Do our teams trust AI answers, or check them all?

Explore what Fathomly can uncover

Everyday use

What's missing between rollout and routine?

One more window

I have to rebuild the whole case in another window.

Read the interview

Fathomly, interview agent: When did you last try the new AI tool at work?
Employee: Yesterday, for a customer reply. I opened it, then went back to our usual template.
Fathomly, interview agent: What happened between opening it and going back?
Employee: The order history was in the ticket. I have to rebuild the whole case in another window. The template was already there.

Possible team decisionThe team could test a reply workflow using the context already in the approved support tool.

Training

Which trainings do our teams really need?

Beyond the demo

I can write the prompt. I can't tell if the result is right.

Read the interview

Fathomly, interview agent: What have you tried from the AI training in your own work?
Employee: I asked it to combine two stock lists. The demo looked easy, but my totals didn't match.
Fathomly, interview agent: What did you do when the totals didn't match?
Employee: I checked the rows by hand. I can write the prompt. I can't tell if the result is right.

Possible team decisionThe team could run a practical session on checking spreadsheet results, using a sample dataset.

Automation ideas

What would our teams automate first?

Chasing changed dates

The work is spotting which dates have changed.

Read the interview

Fathomly, interview agent: Which repetitive task would you most like help with?
Employee: Updating our delivery list. I go through supplier emails every Monday and copy the dates across.
Fathomly, interview agent: Which part of that takes the most attention?
Employee: A new email can replace last week's promise. The work is spotting which dates have changed. Then I tell the people planning around them.

Possible team decisionThe team could trial flagging changed delivery dates for review before updating the shared list.

Sales

Is Sales still writing proposals by hand?

Before the first draft

I need to know which price and promise still apply.

Read the interview

Fathomly, interview agent: How did you put together your last proposal?
Employee: I reused an old one, checked the prices and asked a colleague about the delivery promise.
Fathomly, interview agent: What needed checking before you could reuse it?
Employee: The wording was fine. I need to know which price and promise still apply. They're kept in different places.

Possible team decisionSales could bring current prices and approved commitments together before testing AI drafts.

Trust

Do our teams trust AI answers — or check them all?

Sounds right. Is it?

I can't see which version of the policy it used.

Read the interview

Fathomly, interview agent: How do you decide whether to use an AI answer?
Employee: For travel expenses, I open the policy and check the answer before passing it on.
Fathomly, interview agent: What are you looking for in the policy?
Employee: Whether the limit changed this year. I can't see which version of the policy it used. A confident answer doesn't tell me that.

Possible team decisionThe team could test answers with visible source links and policy dates, keeping review for uncertain cases.

Expectations

Do our people see AI as help or threat?

What happens to the time?

If I'm quicker, does next month's target just go up?

Read the interview

Fathomly, interview agent: How do you feel about using AI in your daily work?
Employee: I'd welcome help with case notes. I'm less sure about all the talk of saving time.
Fathomly, interview agent: What does that talk leave you wondering?
Employee: We already have a full queue. If I'm quicker, does next month's target just go up? Nobody has explained what changes for us.

Possible team decisionThe team could clarify workload expectations and how any time saved would be used before expanding the rollout.

Legal

What does Legal use when ChatGPT is off-limits?

A rule needs a route

We know what we can't use. We don't have an alternative.

Read the interview

Fathomly, interview agent: How do you handle contract comparisons when public AI tools are off-limits?
Employee: I compare the clauses by hand. We were told not to put contracts into ChatGPT.
Fathomly, interview agent: What options are available for that comparison today?
Employee: Word shows the edits, but I still work through their meaning. We know what we can't use. We don't have an alternative.

Possible team decisionLegal and IT could assess a contract-comparison option against their data requirements, starting with sample documents.

Can HR use AI with personnel data in play?

Where is the boundary?

Removing the name doesn't tell me what else is okay.

Read the interview

Fathomly, interview agent: Where have you considered using AI in your HR work?
Employee: For reference letters. I stopped because the examples include details about someone's role and performance.
Fathomly, interview agent: What was unclear when you decided to stop?
Employee: The guidance says no personal data. Removing the name doesn't tell me what else is okay. I'd need an example I can follow.

Possible team decisionHR and the privacy team could define permitted inputs and create fictional examples for the approved workflow.

Tool fit

Which team needs specialist AI, not ChatGPT?

The table matters

It reads the words but loses which value belongs to which part.

Read the interview

Fathomly, interview agent: Where does your current AI tool fall short for your work?
Employee: Comparing supplier specifications. It summarizes the PDFs, but I still rebuild the comparison table.
Fathomly, interview agent: What do you have to correct in that table?
Employee: The scanned columns get mixed up. It reads the words but loses which value belongs to which part.

Possible team decisionThe team could compare tools on representative specification sheets, including table accuracy and review effort.

Engineering

Are our developers using the coding AI?

Installed isn't integrated

It doesn't know the internal helper we already use.

Read the interview

Fathomly, interview agent: How have you used the coding assistant recently?
Employee: For a small test, it helped. For a feature change, I rewrote most of the suggestion.
Fathomly, interview agent: What needed rewriting in the feature change?
Employee: It added a new dependency. It doesn't know the internal helper we already use. The suggestion didn't fit our codebase.

Possible team decisionEngineering could test whether approved project context improves suggestions, keeping normal code review.

Marketing

Does AI copy still sound like our brand?

Polished. But not us

It sounds like a launch announcement, even when we're apologizing.

Read the interview

Fathomly, interview agent: What do you usually change in an AI copy draft?
Employee: The tone. Product posts are okay, but I rewrote the email about a delayed order.
Fathomly, interview agent: What about that email needed a different tone?
Employee: The customer needed a clear explanation. It sounds like a launch announcement, even when we're apologizing.

Possible team decisionMarketing could add approved examples for sensitive situations to the writing guide and review new drafts against them.

Data handling

Is customer data ending up in public AI chats?

More than a name

I hadn't thought about the order number in the message.

Read the interview

Fathomly, interview agent: How do you prepare a customer message for an AI chat? Please leave out any real details.
Employee: For a difficult reply, I sometimes use a public chat. I copy the message and remove the name.
Fathomly, interview agent: How do you decide what to remove?
Employee: I look for names and email addresses. I hadn't thought about the order number in the message. I'm not sure what else counts.

Possible team decisionThe team could clarify data-handling rules with fictional examples and make an approved help route easy to find.

Progress

How far along are we really with AI?

Tried it. Kept it?

I tried plenty. Only the meeting summaries stuck.

Read the interview

Fathomly, interview agent: Which AI uses have become part of your normal week?
Employee: Summarizing my meeting notes. I also tried reports and planning, but don't really use it for those now.
Fathomly, interview agent: What made the meeting summaries worth keeping?
Employee: My notes are already in one place and easy to check. I tried plenty. Only the meeting summaries stuck.

Possible team decisionThe team could separate reported experiments from regular uses and explore why some tasks became routine.

Shared practice

Which teams are moving with AI — and why?

A place to compare notes

We bring the awkward results, not just the good ones.

Read the interview

Fathomly, interview agent: What helps your team use AI in everyday work?
Employee: We spend ten minutes on it in our weekly meeting. Someone brings a draft and we look at it together.
Fathomly, interview agent: What do you discuss when you look at the draft?
Employee: Last week a summary missed a key exception. We bring the awkward results, not just the good ones. Then we update our shared example.

Possible team decisionThe team could offer time for shared practice elsewhere and ask whether it is useful in those teams' work.

Behavioral perspective

Understand what shapes behavior at work

Low usage alone doesn’t explain its cause. Fathomly asks about concrete work and what gets in the way. Different barriers may call for different responses, rather than another training session.

In development: structured behavioral analysis and testing changes over time.

I don’t trust it.

What would you need to check before relying on an answer?

Explore ways to check an answer against its source.

My old process works.

What would take the most effort if you switched?

Explore a manageable first step within the current process.

I hardly use it.

Which task did you last try it on, and what happened?

Explore whether the tool fits the task at all.

The product today

From conversation to next step

Employee input, brought together as themes, evidence, and actions.

Fathomly admin dashboard showing consolidated topics and mention counts for a fictional AI adoption campaign — **Current product UI** with synthetic demo data

PrepareA brief becomes an interview guide.

Generate a guide from a short brief, then share invitation links. No participant login.

ListenAdaptive interviews, 5–20 minutes.

Type or dictate an answer and confirm the transcribed text.

Video calls planned Q1 2027

UnderstandFind themes. Explore the evidence.

Evidence quotes are redacted for personal details. Use Ask to question released campaign evaluations.

Act & learnTrack actions. Start another cycle.

Track action status, make a change, then manually start another interview cycle to hear what people report.

Company-goal reporting with printable evidence and CSV export.

For any company, any size

Three questions companies bring to us

Starting with AI

Your situationYou want to know where AI would really help before you commit budget and licenses.

What Fathomly doesFinds where teams lose time to repetitive work and what training people actually need.

What would our teams automate first?
Which team needs specialist AI, not ChatGPT?
Which trainings do our teams really need?

Rolling out AI

Your situationThe tools are in place, but usage is uneven and you don’t know why.

What Fathomly doesHears what slows people down and what moves the teams already further ahead.

What's missing between rollout and routine?
Which teams are moving with AI — and why?
Is customer data ending up in public AI chats?

AI is rolled out. Is it working?

Your situationThe rollout is done, but no license statistic tells you whether AI has landed in daily work.

What Fathomly doesLets you make one change and check in the next interview round whether it worked.

How far along are we really with AI?
Do our teams trust AI answers — or check them all?
Does AI copy still sound like our brand?

Tell us your question

AI adoption is our first use case. The same approach carries over to other new tools and ways of working.

Built for candid input. Conversations are pseudonymous. In confidential pilot mode, authorized admins may read pseudonymous conversation histories. Group reporting is released for five or more people, never as individual employee scoring.

Each customer has a separate deployment and database.

How a pilot starts

Start with one question that matters

Tell us what you want to understand. A pilot starts within hours, not weeks: no IT project, no integrations, no login for participants.

Your question What do you want to know about your company? A short briefing, one conversation.
The interview guide We build it from your briefing.
The link You share it with your employees through your own channels. No login, no integrations with existing systems.
The interviews 5 to 20 minutes each, typed or dictated.
The findings We review them together: themes, supporting evidence, next steps. Group results from five people.

The first pilot round is free for companies with up to 250 employees, Q3 to Q4 2026.

Or email [email protected]

Understand your people.Make change work

Listen closely.See the bigger picture

Start with the work.

Keep the meaning.

Listen across people.

Make it actionable.

Six interviews. Different perspectives.

Project updates are being prepared more than once.

Is this software worth rolling out?

What’s helping. What’s holding you back

AI opportunities

Tool needs

Everyday AI use

Shadow AI and safety

New tool rollouts

Training and tool value

Explore what Fathomly can uncover

One more window

Beyond the demo

Chasing changed dates

Before the first draft

Sounds right. Is it?

What happens to the time?

A rule needs a route

Where is the boundary?

The table matters

Installed isn't integrated

Polished. But not us

More than a name

Tried it. Kept it?

A place to compare notes

Understand what shapes behavior at work

From conversation to next step

Three questions companies bring to us

Starting with AI

Rolling out AI

AI is rolled out. Is it working?

Start with one question that matters

Thanks for your message

Understand your people.
Make change work

Listen closely.
See the bigger picture