Calmo · Signal Report
2026

Interviews · Nortera · Feb 3 – Mar 11, 2026

The signal
inside Nortera.

We spoke with 47 people across five functions over five weeks. This is what they told us about AI at work — where it's landing, where it's stalling, and what to do in the next quarter.

Prepared for Nortera leadershipScroll
01How we listened

47 conversations, an average of 38 minutes, recorded with consent and transcribed by Calmo.

Interviews were 1-on-1, conducted in English and Finnish, and coded against a shared taxonomy of tasks, tools, emotions, and blockers. No incentives were offered. Participation was self-selected after an all-hands invitation — representative, not statistically weighted.

47 interviewees

Engineering · 16
Product & Design · 9
Operations · 8
Commercial · 7
People & Finance · 7

By tenure

  • < 1 year11
  • 1–3 years18
  • 3–7 years12
  • 7+ years6
02The finding

82%

of people already use AI every week — but only 19% feel confident about when not to.

Adoption is not the problem at Nortera. Judgement is. The people we spoke to are fluent with the tools; what they're missing is a shared sense of where AI belongs in the work and where it doesn't. That gap shows up everywhere else in this report.

19%Confident

"Do you feel confident about when AI is and isn't appropriate?"

Confident about when not to use AI19%
Unsure — depends on the case54%
Not confident at all27%

"I use it every day. I also have no idea if I'm allowed to put a customer's name in it. So I just… paraphrase."

Account manager · 2 years at Nortera
03Where the energy is

Tasks people have already shifted to AI, ranked by how often they came up unprompted.

Writing, summarising, and searching dominate. Numerical analysis lags — not because the tools can't, but because trust isn't there yet.

Drafting & editing long-form writing82%
Summarising meetings and threads71%
Searching internal knowledge64%
Writing or reviewing code58%
Preparing client-facing artefacts44%
Structured data work & analysis31%

Adoption by function

FunctionUses weeklyUses dailyConfident
Engineering
94%
68%24%
Product & Design
89%
52%18%
Commercial
85%
48%14%
Operations
78%
38%22%
People & Finance
62%
24%9%

Weekly and daily usage are high across every function. Confidence is low across every function. The gap is uniform — this is a policy problem, not a training problem.

04Where it stalls

Five blockers came up in more than one in four interviews. Compliance clarity is the largest, by a margin.

01Unclear what's allowed with customer data61%
02Answers drift from Nortera's tone of voice54%
03Switching cost between tools kills flow47%
04No shared prompts — everyone reinvents39%
05Outputs need too much cleanup to trust28%

"Every team has written the same prompt to summarise a call. Thirty times. Nobody shared one."

Ops lead · 6 years at Nortera
05What we recommend

Move 1 · This quarter

Publish a one-page "what's allowed with customer data" note and pin it in every workspace.

The single highest-leverage move. 61% of interviewees named this gap unprompted; clearing it unlocks the other four.

Move 2 · Next 60 days

Stand up a shared prompt library with three house styles — drafting, summarising, internal search.

This addresses the 'reinvent every time' blocker (39%) and the tone-drift blocker (54%) in one move.

Move 3 · This half

Re-interview the same 47 people in Q4. Track the confidence number, not the usage number.

Usage is already high and will only climb. The metric that moves the business is the 19% confidence number — that's what we'd benchmark against.