What Should I Do After Completing AI Literacy Training?

Completing an AI literacy programme is a genuine milestone, but it is not the finish line.

Many financial services firms invest heavily in foundational training, only to find that momentum disappears within weeks. Staff return to their normal workload, the training slides gather dust, and no one has actually used any of the awareness they built.

This is a predictable, avoidable gap. Training builds awareness. Applying that awareness, deliberately and under proper oversight, is what builds fluency.

This article explains what should happen next: how individuals, managers and programme owners can turn literacy training into practical, governed use, and how this connects to broader workforce expectations such as the government's Skills Compact.

Why training completion isn't the end point

Organisations often treat a completed training session as a deliverable in itself. Attendance is logged, a certificate is issued, and the programme is considered done.

But awareness fades quickly without reinforcement. Within a few weeks of an AI literacy session, most people can recall the headline messages but struggle to say what they should actually do differently in their role.

This matters for two reasons. First, it represents wasted investment: firms pay for training time and materials but see little behavioural change. Second, it leaves a gap between what regulators and initiatives like the Skills Compact expect (a genuinely AI-literate workforce) and what actually exists day to day, which is a workforce that once sat through a session.

The operational problem, then, is not the quality of the training itself. It is the absence of a defined bridge between finishing training and using that knowledge in real work.

How organisations have traditionally handled post-training follow-through

This is not a new problem. Learning and development teams have long grappled with how to make any training "stick," and several established approaches apply here.

Refresher sessions are common: short follow-up sessions a few months after the initial training, designed to re-surface key concepts. These help with retention but do little on their own to build practical confidence, because they still tend to be knowledge-based rather than task-based.

Manager check-ins are another traditional mechanism. A line manager asks, during a regular one-to-one, whether an employee has had a chance to apply recent training. This can be effective, but only if managers are given a specific prompt or opportunity to discuss. A generic "have you used any AI lately?" rarely produces a useful answer.

Competency frameworks offer a more structured route. Some firms define tiered competency levels (aware, capable, proficient) and expect staff to progress through them with evidence. This works well for technical or regulatory skills with clear assessment criteria, but AI literacy is newer and less standardised, so many firms have not yet built this scaffolding.

All three approaches remain valid and useful. The gap in most organisations is not a lack of mechanisms, but a failure to apply any of them specifically to AI literacy, treating it instead as a one-off compliance exercise.

Where AI itself helps reinforce learning

One of the more effective ways to convert literacy into fluency is to use AI tools, under supervision, as part of ordinary low-stakes work.

This might mean using an AI assistant to draft a first version of a routine internal summary, to research background context on a topic before a meeting, or to generate a first pass at reformatting data for review. In each case, the AI produces a starting point that a person then checks, edits and takes ownership of.

The value here is practical, not theoretical. Reading about how large language models work is useful, but it does not build the same confidence as actually using a tool, seeing where it performs well, and noticing where it gets things wrong. Repeated, low-stakes exposure like this is what shifts someone from "I attended a session on AI" to "I know what this tool is reliable for, and what it isn't."

This only works safely if oversight is built in from the start. Every output should be reviewed by a competent person before it is used or relied upon, exactly as it would be if a junior colleague had produced the first draft. AI accelerates the drafting stage. It does not remove the need for someone experienced to check the result.

What to put in place operationally

Turning literacy into applied capability requires a small number of concrete operational steps.

Individuals should identify one or two specific, low-risk tasks in their own role where AI awareness could be usefully applied, rather than waiting for a broad instruction to "start using AI." A specific task, such as summarising a long document for internal review, is far easier to act on than a general aspiration.

Managers should actively sponsor these opportunities. This means allocating time for a team member to try something new, agreeing what oversight will look like, and following up on how it went. Without this active sponsorship, most staff default back to old habits under normal workload pressure.

Governance does not pause once training ends. Existing controls around data handling, client confidentiality, and sign-off on client-facing material continue to apply in full. If anything, the period immediately after training is when oversight checkpoints matter most, since staff are experimenting with new tools and workflows for the first time.

Finally, progress should be tracked, even simply. Recording which teams have moved from training completion to actual, supervised application gives the organisation a much stronger evidence base for demonstrating Skills Compact alignment than attendance records alone.

Example

A mid-sized London specialty insurer completes a firm-wide AI literacy programme for its underwriting and claims teams.

Three months later, an underwriting team lead notices that no one has actually used any AI tool since the sessions ended. She works with the learning and development lead to identify a specific, low-risk task: using an AI assistant to summarise long claims correspondence for internal review, with all outputs checked by a senior handler before use.

Within a few weeks, the team reports increased confidence using AI tools appropriately, with oversight checkpoints preventing any unchecked output from reaching clients. The firm uses this as a template for rolling out similar low-risk applications across other teams.

FAQs

• Do I need further training before I start using AI in my role?

  Usually not. Foundational literacy training is generally sufficient to begin applying AI in low-risk tasks, provided your organisation's existing oversight and governance controls remain in place. Further, more technical training can follow once specific needs become clear through practical use.

• Who is responsible for identifying what I should do next?

  It is a shared responsibility. The individual is best placed to spot relevant tasks in their own role, the manager should actively create space and sponsorship for trying them, and the programme owner should ensure there is a consistent approach and appropriate oversight across teams.

• How does this relate to the government's Skills Compact?

  The Skills Compact is concerned with building a genuinely AI-capable workforce, not simply one that has attended training. Demonstrating ongoing, supervised application of AI skills is increasingly relevant evidence of that alignment, alongside training completion records.