Fund workforce capability after Slack’s skills update

Put AI workforce enablement strategy behind the budget

Your AI workforce enablement strategy needs a clear answer to a practical question: which parts of better work should live in reusable instructions, and which require employees to build judgment through practice? We recommend funding both, with separate evidence for each.

On 28 October 2026, Slack announced and shipped an improved Skills experience, including discovery, administrator verification and team assignment. These skills package repeatable instructions for Slackbot. The release also included event-triggered tasks and Slackbot Mini, bringing assistance alongside other applications. For learning leaders in MENA organisations already using Slack, this creates a reason to review how adoption budgets connect learning with everyday work. Slack announcement

Slack describes Enterprise+ launch access without limits for Advanced Features, while reserving the possibility of future limits with 30 days’ notice. Before funding a rollout, obtain written confirmation of your tenant’s feature entitlement and regional access, alongside the applicable commercial terms. Slack availability statement

Separate the investments you are approving

We would divide the business case into three allocations. Each needs an owner, an output and a recurring budget.

Allocation	What you fund	Evidence you require
Employee capability	Practice, coaching and assessment	Employees can judge, correct and explain an AI-assisted result
Reusable work guidance	Approved instructions, examples and maintenance	The guidance produces acceptable drafts across representative cases
Operating ownership	Content review, support and change management	A named team can update or withdraw guidance when procedures change

This division helps you spot an incomplete proposal. A library of instructions needs people who understand when to use it. A training programme needs a route for useful practices to reach daily work. Both need maintenance after launch.

Ask your sponsor to name the business outcome before approving any allocation. “More AI engagement” leaves too much room for activity without value. “New supervisors can prepare an accurate escalation brief using approved guidance” gives learning, operations and technology teams a shared target.

Choose one role and one recurring task

Your enterprise AI adoption roadmap should start with a role whose work can be assessed safely and consistently. We suggest a six-week pilot with approximately 20 participants, subject to your staffing and risk constraints.

A suitable example is preparing an internal handover brief from approved, fictional operational notes. The task should have a clear audience, a recognised quality standard and an accountable reviewer. Avoid starting with decisions about employee performance, customer eligibility or other sensitive outcomes.

Write a one-page scope containing:

• The role, task and business owner.
• The approved inputs and prohibited information.
• The expected output and its reviewer.
• The employee judgment that remains necessary.
• The evidence required to expand, revise or stop.

For a UAE or Saudi team working in Arabic and English, include representative examples in both languages if both are required. Ask local reviewers to assess terminology, tone and missing context separately. A strong English result should not decide the Arabic acceptance outcome.

Test performance and learning separately

Our proposed workforce AI pilot design has three stages.

During the first week, collect a baseline using the existing process. Record preparation time, reviewer corrections and the participant’s explanation of important decisions. Use the same quality rubric throughout the pilot.

During weeks two to four, introduce one approved skill and a short practice session. Give participants examples of an acceptable draft, a misleading draft and an incomplete input. Ask them to identify the problem before correcting it. Keep a log of support requests and changes to the instructions.

During weeks five and six, assess comparable tasks under two conditions: with the approved assistance and without it. Use these assessments to examine task execution with assistance and employees’ ability to apply learning independently.

Set acceptance conditions before participants begin. We recommend requiring:

• Every assessed output to receive the designated human review before use.
• No unresolved critical error in the final evaluated outputs.
• Participants to recognise deliberately incomplete or conflicting inputs.
• Separate quality approval for each required language.
• A successful exercise in updating or withdrawing obsolete guidance.

Your operations owner should define a critical error. Report results by condition and language, including failures and support effort. Agree the acceptable time and quality thresholds with that owner before seeing results.

Cost the capability through its useful life

Compare proposals using the same scope and period. We suggest calculating:

Pilot cost = incremental platform charges + employee time + reviewer time + content preparation + integration and administration + evaluation.

Then estimate the recurring cost of maintaining each approved skill. Include subject-matter review, language updates, support, reassessment and changes caused by new procedures.

Divide the total by the number of employees who meet your acceptance conditions to estimate cost per demonstrated capable employee. Keep this separate from cost per person invited or completing a session.

Also record cost per accepted work output. An accepted output passes your rubric after review. Include correction time so that faster drafting does not obscure additional review effort. Treat any estimated annual benefit as a planning assumption until repeated operational evidence supports it.

Resolve data boundaries before connecting work sources

Ask the supplier and your technology team to document what information enters the service, where it is processed, how long it is retained and who can access it. Include prompts, attachments, outputs and usage records. Start with synthetic material while these questions are resolved.

For UAE operations, ask counsel to check the applicable framework, including the UAE federal personal data protection law, Federal Decree-Law No. 45 of 2021. Official US government overview

For Saudi operations, ask counsel to check Saudi Arabia’s Personal Data Protection Law and its Implementing Regulations. Saudi Central Bank reference

Require the appropriate internal owners to approve the proposed data boundary before introducing real work records.

Buy a decision you can defend

Your AI enablement provider selection should require a pilot scope, responsibility map, complete cost assumptions, assessment rubric and handover plan. Ask who owns the reusable guidance and who funds its maintenance. Make expansion conditional on demonstrated capability, acceptable work quality and sustainable operating effort.

Through DeepStrategy, we work with leadership on AI investment and transformation strategy. Our process starts with discovery and readiness assessment, then defines a roadmap and success measures aligned with business objectives.

Bring your target role, existing learning budget and one recurring task to a conversation with us. We can discuss the investment decision and the evidence your leadership team needs before committing to a wider programme.

Sources

1. Slackbot: Delegate more work, in more places (2026-10-28)
2. United Arab Emirates Allows Cross Border Data Flows of Personal Data (2024-01-22)
3. Compliance with the Updated Personal Data Protection Law and its Implementing Regulations (2023-09-18)
4. DeepSolve Our Divisions; undated, accessed 6 November 2026 (2026-11-06)
5. DeepSolve Process and Tech Stack; undated, accessed 6 November 2026 (2026-11-06)
6. DeepSolve homepage and contact form; undated, accessed 6 November 2026 (2026-11-06)

Prepared with AI assistance. DeepSolve is responsible for the published content. Illustrations are conceptual.