Most companies treat internal mobility as a matching job. Each employee has a set of skills. Each role lists the skills it needs. A mobility platform compares the two and shows who could move where.
The match is only as good as the skills listed on each role, and someone has to keep those lists up to date. Jobs keep changing, so a list that was right two years ago can point people toward the wrong skills today.
Keeping the lists current is hard at scale. A large company can have thousands of roles, each listing a dozen or more skills, and only a few people to look after them. A small team in your organization will struggle to keep thousands of roles current by checking them one by one. It needs help to find which changes matter most and how each one fits the roles around it.
To do this, your organization needs:
Together, they keep the skills listed on each role in step with how the job is done today. Your mobility platform then matches people against what each role really needs, not what it needed when the list was last updated.
The harder question is why the skills listed on each role fall out of date so easily.
Role skill lists fall out of date for three reasons. AI is changing jobs faster than the lists are updated. Busy teams update one role at a time and leave the roles around it behind. Small teams struggle to check thousands of roles one by one. Each reason leaves people preparing for jobs as they used to be, so each one makes internal moves harder to plan.
Each role has a list of the skills it needs. That list is only as fresh as the last time someone checked it. AI is making lists go stale faster. AI tools now do some tasks and change how people do others, so jobs need new skills and drop old ones. Lightcast, a company that studies job postings, found that the average US job changed 32% of its required skills between 2021 and 2024. It said AI drove more of that change than anything else.
When the list is out of date, mobility stalls in three ways. Employees learn the skills the role used to need, so they are not ready when a move opens up. People who already have the new skills do not see them asked for, so they do not apply. Managers see internal people fall short and may start hiring from outside instead.
Roles on the same career path often share a skill, but each role may need it at a different proficiency level. A more senior role often needs more of the skill, but not always. A manager might need less hands-on skill in a tool than the analyst who uses it every day. Before applying for the next role, an employee compares it with their own to see what they still need to learn.
Skills teams are busy, so they often update only the role someone asks about. They add a skill to one senior role at a high level and leave the roles below it alone. The career path now has a missing step. The senior role asks for the skill, and the role below never mentions it.
That missing step stalls mobility. Employees in the lower role have no reason to learn the skill, because their role never asks for it. When they look at the next role, they find they need it at a high level and are starting from scratch. Some decide the move is out of reach and stop trying. Others apply and are turned down for a skill they never knew they needed.
Redeployment stalls the same way. When a company moves people out of shrinking jobs, it trains them for the skills their new role lists. If the path above that role has missing steps, the training prepares them for the new job but not for the next role up. People settle into the new role, then find the next move harder to reach than it looked.
A large company can have thousands of roles. Each role lists a dozen or more skills. Only a few people keep all those lists up to date, and they need help from managers who know each job. With so little time, they can check only a small share of the roles each year.
Checking one role at a time also wastes time. The same skill might be missing from twenty roles, and the team finds it and fixes it twenty separate times. The team also cannot easily tell which fixes matter most. A fix to three roles that employ hundreds of people matters more than a fix to twenty roles that employ a few dozen.
Mobility stalls in the roles nobody checks. Their skill lists stay out of date until the next review. AI keeps changing those jobs in the meantime, so the lists fall further behind. The people in those roles keep learning skills the job no longer needs.
A skills team has two questions to answer. Which fix should it make first? Does that fix still fit the roles around it?
Fuel50's Ontology Engine and Curation Queue answer both questions, and they work as a pair.
The Ontology Engine does the research. It compares each of your roles with similar jobs in the external job market and spots skills to add, skills to check and the proficiency level each role needs.
The Curation Queue is where your skills team acts on that research. It sits inside Skills Inventory, turns the engine's suggestions into a short list of decisions and sends each approved decision through your usual approval process.
A suggestion moves from the market to your roles in six steps.
The engine matches each of your roles to similar jobs in the market.
It compares the skills those jobs ask for with the skills your role lists.
It suggests skills to add or remove, and a proficiency level for each role.
The queue groups the suggestions by skill and puts first the changes that affect the most people and have the highest impact across your organization.
Your team checks the roles, proficiency levels, functions and business units involved.
Approved changes go through the Skills Inventory approval process before they update any role.
Here is what your team can do with those suggestions in the Curation Queue.
Without the queue, a skills team starts from one role and works through its list of skills. The queue turns that around. It starts from one skill and shows every role where that skill is suggested. One item shows those roles, the proficiency level suggested for each, how many employees work in them, and which functions and business units they sit in. Your team looks at the skill once, then decides role by role which changes to keep.
The Curation Queue surfaces the changes that affect the most people and ranks them by their impact across your organization, so the highest-impact changes come first. It also shows how many roles each item touches, but the number of roles does not set the order. A skill suggested for three roles can come before one suggested for twenty if more people work in those three roles. That points a small team to the fixes that matter most to the organization and to the employees looking for their next move. The order tells your team where to start. Your team still decides what to approve.
Often the skill already exists in your company's skills list but is missing from roles that need it. The queue shows which roles, and the proficiency level suggested for each. Your team checks that the skill fits the work in each role and sends the change for approval. The admin who approves it can choose which roles get the skill and at what proficiency level. The skill is added to those roles without being created again.
Sometimes similar jobs in the market ask for a skill your company has never listed. Your team can start creating that skill from the same queue item. It checks the skill's definition and whether it matters to your business, then decides which roles need it and at what proficiency level. Those role changes then go through approval like any other.
For each group of related roles, the engine suggests the proficiency level each role needs. Levels often rise with seniority, but not always. A senior role can need a lower level of a skill than a junior role that uses it every day. The engine builds these groups of related roles from market data, not from the career tracks you have already set up. Reviewers look at the whole path instead of copying one change everywhere, and change any level that does not fit how your business works.
Now and then the engine flags a skill on your role that similar jobs rarely ask for. These suggestions are rare. They prompt your team to check whether the skill still matters to your business, because a skill that is rare in the market can still be key to how you work. If your team does remove a skill from some roles, the skill stays in your company's skills list.
The market shows what similar jobs ask for. It cannot know your strategy, and matching your roles to market jobs is never exact. Every suggestion therefore goes through people before it changes a role. A team member sends a suggestion for approval or turns it down. A Skills Inventory admin then reviews it, can change the roles and proficiency levels, and approves or declines it.
A queue gives your team better suggestions and a clearer place to start. The value still depends on how your team runs its reviews, who checks each change and how you measure progress.
Start by answering four questions about how your team keeps role skills current today.
The first question you answer no to is where to start. A no to either of the first two means your team needs a way to see patterns across roles. A no to either of the last two means it needs clearer rules for how changes are checked and approved.
Market data can show what similar jobs ask for. Only someone who knows the work can say whether a skill fits your business. For each group of related roles, name a reviewer from the business, such as a sales leader for sales roles.
Agree up front what that reviewer checks. They confirm whether the skill is needed, whether the proficiency level is right for each role, and whether a removal suggestion should be acted on or ignored. Then agree who sends changes for approval and who gives final approval in Skills Inventory, so no suggestion sits in the queue without an owner.
The engine's suggested proficiency levels reflect what similar jobs in the market ask for. Your roles may need different levels, and a senior role will not always need more of a skill than a junior one. Check each suggested level against what people in that role actually do, and change it where your business works differently. Apply only the changes that fit your organization.
Decide how often your team reviews role skills and how many changes it can handle each time. Shorter, regular reviews make it easier to keep up as AI changes jobs, because each review deals with fewer changes at once.
Within each review, start with the changes that affect the most people and have the highest impact, then check how each one fits the roles around it. If part of the business is restructuring or moving people into new roles, review those roles first. People need accurate skill lists before they move, so their training matches the roles they are moving into.
Track a few numbers each cycle. Count how many changes your team reviews and approves, how many roles and employees those changes cover, how long it takes from opening a review to the first approved change, and how many broken career steps your team finds and fixes.
Compare the numbers from one cycle to the next. If coverage grows while review time stays the same, your team is keeping up with the jobs. If the same gaps keep coming back, the review rhythm or the reviewer setup needs another look.
Skill lists that keep up with the jobs let people prepare for a move, whether they choose it or the business asks them to make it.
Review your skills queue to see where your role skills need attention first.