Let's Connect

Enterprise AI readiness assessment

Enterprise AI
Readiness
Assessment

Find out what needs to change before you put AI into a workflow. Check the work, the information and the people around it, then leave with a practical next step.

Free to use. No email needed.

Find the real obstacle

What needs to change before AI will help?

A new tool can leave the old problem in place. Start by looking at how the work gets done today. The useful outcome may be a cleaner source document, a simpler handoff or better access. Sometimes it is a small AI test. Sometimes the existing process needs fixing first.

Before buying a tool

Identify the task, the expected improvement and the simpler alternatives.

When a pilot gets stuck

Separate missing access, poor sources, review gaps and skill needs.

Before expanding

Check what changes for the next team, system or type of information.

A practical review

One workflow. Eight checks.

Use these questions with the people who do the work. Open each card for evidence to look for and a possible next action.

01

A problem worth solving

What is slow, unreliable or expensive today?

What to look for

Evidence: Observe the current task. Keep a few examples and record the effort, including review and rework.

A useful next action: Compare AI with a simpler change, such as a standard reply, a better form or a clearer process.

Who can help: The person responsible for the work

02

Sources people can trust

Are the instructions and records current and consistent?

What to look for

Evidence: Check a few real source documents for conflicting instructions, missing dates and unclear ownership.

A useful next action: Choose the current source and fix conflicting versions before using them to generate answers.

Who can help: The document or data owner

03

Clear information rules

Which information can this workflow use?

What to look for

Evidence: Check the rules for the chosen tool and task. Include access, sharing, retention and any sensitive records.

A useful next action: Get an answer for unresolved information questions. Use made-up or suitable public material while you investigate.

Who can help: The relevant information, IT or security owner

04

Access that works

Can the people doing the work use the required features?

What to look for

Evidence: Try the intended steps with their accounts, permissions and files. A successful login is only the start.

A useful next action: Fix the missing permission or feature, or narrow the workflow to what is available.

Who can help: The tool or system administrator

05

A sensible place in the process

What happens before and after the AI step?

What to look for

Evidence: Walk through the handoffs, systems and exception cases. Identify any step that sends, changes or approves something.

A useful next action: Start with a manual handoff where that is enough. Define and test any integration or automatic action separately.

Who can help: The workflow and system owners

06

A way to check the result

How will someone notice a plausible but wrong answer?

What to look for

Evidence: Try typical cases, difficult cases and a case with a known mistake. Name what a reviewer should check.

A useful next action: Create a short review checklist and decide when the work needs a specialist or should stop.

Who can help: The person accountable for the output

07

People who can do the task

Can users explain the result and correct a mistake?

What to look for

Evidence: Ask someone to complete the task and talk through their choices. Note where they need help.

A useful next action: Use focused practice for the gaps you observe. Fix access or source problems rather than teaching around them.

Who can help: The team lead and people doing the work

08

Someone to keep it working

Who owns changes, questions and the next review?

What to look for

Evidence: Name the workflow owner, support contact and review date. Check that they have accepted the work.

A useful next action: Agree who updates the material, handles exceptions and decides whether to continue.

Who can help: The workflow owner and sponsor

How to use the worksheet

Turn a discussion into an action list

Work through one task from start to finish. Write down what you know, what you still need to find out and who can act. You can use the worksheet yourselves or bring it into a conversation with Ajaia.

  1. 01

    Choose the work

    Name the input, the result and the person responsible. Keep a current example for comparison.

  2. 02

    Check the evidence

    Try the access, read the sources and inspect a result. Record a gap instead of guessing.

  3. 03

    Choose the next change

    Give the action an owner and a completion check. Agree when to review it.

Use it now

Your AI readiness worksheet

Choose one workflow. Mark each area using evidence you have actually seen. Leave an unanswered question as “Unknown”. A total score would hide which gap needs attention.

01 / A problem worth solving

What is slow, unreliable or expensive today?

Add evidence and next actions
02 / Sources people can trust

Are the instructions and records current and consistent?

Add evidence and next actions
03 / Clear information rules

Which information can this workflow use?

Add evidence and next actions
04 / Access that works

Can the people doing the work use the required features?

Add evidence and next actions
05 / A sensible place in the process

What happens before and after the AI step?

Add evidence and next actions
06 / A way to check the result

How will someone notice a plausible but wrong answer?

Add evidence and next actions
07 / People who can do the task

Can users explain the result and correct a mistake?

Add evidence and next actions
08 / Someone to keep it working

Who owns changes, questions and the next review?

Add evidence and next actions

Use the evidence to choose what needs to happen next. This worksheet does not approve a deployment.

Your notes stay in this tab. They are not automatically saved or sent to Ajaia. Download them before leaving. Use general descriptions rather than private records.

What to do with the findings

Choose the next step for this workflow

A “Ready” label needs a reason. “Unknown” is a useful finding too: it tells you what to investigate before making a bigger commitment.

Keep the decision specific to this workflow. A working test with made-up information does not establish that a production integration is ready.

Fix the source or process

Resolve conflicting instructions, remove duplicate steps or try a standard template. Check whether that solves the problem.

Resolve access or information questions

Ask the owner for a decision. Change the scope if the required access or information is unavailable.

Run a small comparison test

Compare similar tasks with and without AI. Include checking, corrections, support and cost, as well as speed.

Build or teach what is missing

Scope an integration when the handoff needs engineering. Use focused practice when people need a skill. Check the result before expanding.

A worked example

Should this support team use AI for replies?

Invented example. The team, findings and actions below are made up.

The situation

A support team searches help articles and rewrites similar answers. It wants AI to draft replies, with a person checking each one before sending. Three reviewers can use the approved workspace. Two help articles give different return instructions, and the team has not agreed whether real customer messages may be used.

The decision

Fix the conflicting articles and create a shared reply checklist first. Then compare standard replies with AI-assisted drafts using made-up questions. Keep real customer messages out until the information owner has agreed their use.

The first actions improve the source material and the review process. Buying more licenses or scheduling a general training session would leave those gaps unresolved.

A problem worth solvingUnknown
Finding
The team expects faster replies but has not measured searching, drafting and checking.
Next action
Time a few comparable replies using the current process, standard replies and AI-assisted drafts.
Suggested owner
Support lead
Completion check
The comparison includes review time, corrections and answer quality.
Sources people can trustNeeds work
Finding
Two help articles give different return instructions.
Next action
Choose the current instruction and correct the conflicting article.
Suggested owner
Help-content owner
Completion check
Both articles point to the same approved instruction.
Clear information rulesUnknown
Finding
The team does not have an agreed answer about using real customer messages.
Next action
Ask for a decision about the proposed information use. Keep the first comparison test made up.
Suggested owner
Information owner
Completion check
The decision and any limits are recorded for the team.
Access that worksReady
Finding
All three reviewers completed the intended drafting steps in the approved workspace.
Next action
Recheck access if the test expands to more people or different features.
Suggested owner
Tool administrator
Completion check
Each new participant can complete the same steps.
A sensible place in the processReady
Finding
People will copy drafts into the support system and check them before sending. No automatic replies are proposed.
Next action
Keep the manual handoff during the comparison test.
Suggested owner
Support lead
Completion check
Every outgoing reply still has a human reviewer.
A way to check the resultNeeds work
Finding
Reviewers currently rely on their own judgement without a shared checklist.
Next action
Agree checks for the source, return conditions and unsupported promises. Try a draft with a known error.
Suggested owner
Quality reviewer
Completion check
Reviewers find the known error and explain the correction.
People who can do the taskNeeds work
Finding
One reviewer missed an unsupported promise in a practice draft.
Next action
Practise checking and correcting another draft before the test.
Suggested owner
Team lead
Completion check
The reviewer catches the same type of mistake on a different question.
Someone to keep it workingUnknown
Finding
Nobody has accepted responsibility for questions and source updates after the test.
Next action
Assign a workflow owner and agree the next review.
Suggested owner
Support manager
Completion check
The owner accepts the role and the review is scheduled.

When you want help

Work through the next decision with Ajaia

Bring a short description of the workflow and the gaps you have found. We can review the work with your team and help turn the findings into a prioritised plan.

The work may involve data and source cleanup, process design, tool setup, engineering or targeted practice. We agree the scope and deliverables after a conversation.

Useful preparation

  • The task and the result you want.
  • An example of the current process, using made-up material where needed.
  • The tools and systems involved.
  • The open questions and people who can answer them.

Keep private records out of the inquiry form.

Customer references available upon request.

Tell us what kind of workflow or project you want to hear about.

Practical questions

Questions about AI readiness

Can I use the worksheet without contacting Ajaia?

Yes. Fill it in on this page or download the blank CSV and PDF. The worked example is also public. The on-page worksheet does not automatically save or send your notes; download them before leaving.

Should we assess the whole organisation at once?

Start with a workflow and the people who own it. An organisation-wide view can then compare the findings across workflows, teams and systems. One ready task does not establish that every proposed use is ready.

What if our main problem is poor data or a broken process?

Record that as the next piece of work. Clearer source material, a permission change, a standard reply or a redesigned handoff may help before training or an AI build. Compare the simpler option too.

Does a completed worksheet approve an AI deployment?

No. It records findings and next actions. The relevant owners still decide permissions, information use and deployment. It does not replace a security review or a specialist audit.

How should we measure whether AI helps?

Compare similar tasks using the current process and the proposed approach. Include checking, corrections, support and any new costs. Record the quality of the result as well as time spent, and keep estimates separate from observations.

What does Ajaia help with?

We can review the workflow with your team, investigate gaps and turn the findings into a prioritised plan. The follow-on may involve source cleanup, process design, tool setup, engineering or targeted practice. We agree the scope and deliverables after a conversation.

Can we start without sharing private records?

Yes. Explain the task with made-up or suitable public material first. Keep private records out of the inquiry form. Agree any later access and information-handling arrangements before sharing real work material.

Can we ask for customer references?

Customer references available upon request. Tell us what kind of workflow or project you want to hear about. We arrange references individually.