Visibility into AI adoption

You invested in AI.Now knowwhat it returns.

sonar supports your employees on WhatsApp: it teaches, answers and removes blockers. In the same conversations it learns what is improving, what is blocked and what is needed, by process and department.

On the surface: what management sees today
1,284Logins this month
9,410Messages sent
200Licenses assigned

Activity. Not value.

Beneath the surface

Every point is a person in a work process. sonar hears what is really happening there.

ProgressingBlockedNeeds helpNot known yet

Four answers, by process and department.

What is improvingProcesses that are moving, what changed in them and how much time came back
What is blockingReported blockers: access, permissions, habit
What is neededConcrete requests for leadership
Where it's quietWho hasn't started yet, and why
How it works

Helping employees is also the measurement. sonar asks what changed, inside the help conversation itself.

Step 1

Help on WhatsApp

Answers, resurfaces training materials and sharpens prompts, in the channel people already use.

Step 2

A question in the same thread

What did you try? What improved? What is stopping you? No forms.

Step 3

A picture for leadership

What is improving, what is blocking and what is needed, by process and department.

Management viewIllustrative data
Progress

What is improving

5 / 12 processes

Reported progress, not measured outcome

Focus

What is blocking

Progressing5
Blocked2
Support only2
No update3
Action

What is needed

!Grant access to the expenses folderFinance · raised in a conversation this week
+Schedule practice time for the shift teamOperations
!Define who approves drafts before sendingService
Method

How a WhatsApp sentence becomes a number. Every number traces back to the sentence it came from.

Value · illustrative scenario

It starts with one person.

40minutes a day back to one employee

Fewer reports, less copying between systems, less searching.

Now across the whole organization.

1,400active users out of 2,000 employees

933 hours a day · 40 min × 1,400 users

The time adds up on its own.

224,000work hours returned this year

933 hours a day × 240 workdays · about 110 full-time roles

Hours turn into money.

$19.3Mvalue per year
Time returned$13.4M224,000 hours × $60
New work$3.4MA quarter of the hours went to more proposals, customers and projects · $60 extra value per hour
The stars$2.4M70 users building what we once only dreamed of × $3,000 a month
Licenses reallocated$102,857300 unused licenses × $30 a month

On every dollar spent on licenses.

28×return on license cost

$19.3M a year against $0.7M in licenses · $30 per user per month

Time freed is not necessarily money saved. sonar shows how much time came back and where it went, so you know what you can rely on.

“I can see it on their faces.”What managers tell us after a few weeks

More energyNo changeStill frustrated

896employees report less manual work, and their energy to create is back.

“Finally, I'm not copying data between systems.”“I enjoy the creative part of my job again.”

Illustrative scenario · 2,000 employees · 70% active · $60 an hour · new work is worth $120 an hour, only the extra $60 is counted · 5% stars · $30 license per user per month · aggregate data only

Live · today in the organizationIllustrative simulation · aggregate data only
0work minutes returned today
0people working with AI right now
0processes mapped
0open requests for leadership
Who is behind sonar

Harari Partners

A partnership that guides leadership teams and their people through AI adoption. sonar grew out of that work, so that in every organization we support, we know what actually works.

50+Organizations advised
300+Workshops and trainings
From the field · organization with about 30 participants

Leadership wanted to know which employees needed more capable tools, and why.Instead of a survey, sonar asked inside the help conversations,and the answers drove how subscriptions were allocated across departments.

~5questions reached a human advisor since launch. sonar handled everything else.

Some of the organizations we have worked with
  • Tambour
  • HP
  • KLA
  • Camtek
  • Ormat New Ventures
  • Elbit Systems
  • Vistage
  • AHAVA
  • Maccabi Healthcare Services
  • Harel
  • Gett
  • Reichman University
  • Israel Innovation Authority
  • Israel National Digital Agency
  • Civil Service Commission
  • Manufacturers Association of Israel
  • Electra
  • Granot Group
  • Hever
  • Buligo
  • CBC Israel
  • Lubinski Trade
  • Sphera Funds Management
  • Vargus
  • SCM C.O.O.
  • Ogen
Privacy and trust

People speak freely only when they trust the channel.

The data serves improvement, not surveillance.

What leadership sees

Processes and departments, in aggregate

Procurement · Preparing quotesProgressing · 3 participants
Finance · Expense approvalBlocked: permissions
Operations · Shift reportProgressing · 9 participants

A group is shown only when it has at least 5 participants.

What management never sees

Who said it, and exactly what

Dana Levi054-0000000Honestly I'm not sure my manager knows I'm using this
  1. Transparent from the first message

    Every employee knows they are talking to an AI agent, and what is collected and why.

  2. No view of an individual

    Management sees processes and departments, and only in groups of 5 or more.

  3. Limited retention

    Conversations are kept for 90 days, then deleted. Data is stored in the EU.

  4. Right to know and delete

    Any employee can ask what is stored about them, stop, and request deletion, with no consequences.

  5. Data processing agreement

    A DPA is ready for every customer, available on request, and models don't train on your data.

  6. Employee representative pack

    A ready explanation for works councils and employee reps: what is collected, what is shown, and to whom.

Getting started

A 12-week pilot, with an answer at the end. One group, one owner on your side, and every process measured against a baseline.

Week 0

Setup

Program materials, participant list and a baseline for every process.

Weeks 1-6

Support on WhatsApp

sonar helps employees and learns what is improving, blocking and needed.

Week 6

First picture

A first management view, and first requests for leadership.

Week 12

Value readout and decision

What changed in time and frustration, and the recommendation going forward.

You commit

Little, but clear

  1. An executive sponsor
  2. An adoption lead and one or two departments, up to 40 participants
  3. 30 minutes a week to sync
You get

A picture you can rely on

  1. Ongoing support for employees on WhatsApp
  2. A management view: what is improving, blocking, needed
  3. A baseline and a value readout for every process
If the readout isn't useful, you walk away.If you continue, you pay only for active participants, and the pilot fee is credited.

See where AI
is working for you.

20 minutes: a real WhatsApp conversation, and the management view built from it.

Book 20 minutes
  • A live demo, not slides
  • What a pilot would measure for you
  • No commitment