Greater efficiency means little if L&D cannot demonstrate its contribution to business success. Ryan Austin explores how smarter capacity planning, meaningful measurement and responsible AI adoption can transform learning operations, helping teams prioritise strategic work, make better decisions and build the credibility needed to influence business leaders at every level.
Part one of this feature covered:
- The pressure in the numbers: in 2026, 12% fewer people logged the same hours, the busiest 10% took on 65% of the work, and more than a third of active projects had no priority
- Build one view of the function: a single governed record of requests, projects, tasks and hours, instead of rebuilding the answer by hand every quarter
- Measure the backlog before you try to shrink it: track unprioritised work as a standing metric and tie every project to a company goal
- Turn intake into a filter: score every request on goal, consequence and owner, and check if it is really a training problem before you build
- Govern the work without adding paperwork: a federated model, clear decision rights and governance built into the tools people already use
And now, continuing with what to do once you can see the work:
5. Treat capacity as a warning light
The clearest number in our research came from a CLO who leads 53 people serving about 15,000 employees. I asked what he would want to see on a Friday afternoon. He answered right away: how his team’s capacity is tracking against five strategic priorities versus reactive work.
His estimate today is that about 10% of capacity goes to strategy and 80 to 90% goes to “answering the mail on what’s already in flight.” His goal is 70/30.
Our platform data shows why that matters. Across the functions we studied, the number of people actively logging work fell by about 12% in 2026, while total hours held flat. The same work moved through fewer hands. The share carried by the busiest 10% of contributors rose from 49% to 65%.
Call that what it is: a concentration problem, not an efficiency win. It is one of the more reliable early warnings of attrition. When it lands, it shows up as an urgent, expensive backfill instead of a planned hire. And it is exactly what a finance team should be watching, well before it turns into a resignation letter or a missed launch date.
“Sometimes when we have really cool projects, it’s like a bunch of flies and honey. Everyone just kind of grabs that project and I look up with seven people on one project.”
Senior Director of Learning, environmental services
Capacity is necessary, but it is not enough on its own. Producing more content in less time can still mean more low-value output. Freed time earns L&D a credible seat in a budget conversation. What you do with that time decides whether you keep it.
Actionable steps:
- Set a target ratio for strategic versus reactive work that fits your function, such as moving from 10/90 toward 70/30. Review it monthly and show it to leadership
- Report workload concentration as a governance metric, not a compliment. If your busiest 10% carry two-thirds of the work, model the cost of replacing even one of them. Put that number on the same slide as any headcount request
- Separate planned delivery from support, admin and unplanned demand. One hour of design time is not the same as one hour of LMS admin
- Don’t aim for 100% utilisation. A team with no slack can’t absorb change or fix quality issues
- When you free up capacity, say where it goes: faster response on top priorities, deeper diagnosis, better quality checks or lower vendor spend. Otherwise “hours saved” just turns into more low-priority work
Tip: Solve prioritisation before capacity. As one leader put it in a live session, they are two problems solved in order. First decide what matters. Then make room for it.
6. Make two promises, in order
As L&D moves under the CHRO in more organisations, the bar for showing value goes up. One CLO named the core problem: teams optimise for specific requests instead of business goals. They answer the wrong question, for the wrong audience.
“If you can correlate that the work you’re doing on the learning and talent agenda is driving the company goal, they’re never going to get rid of you. But if you keep them two separate, when there’s cost pressure, there’s usually more focus on those teams.”
Head of Sales Enablement and Training, healthcare data and analytics
Leaders who have earned executive trust tend to make two promises, in order. A former L&D leader told his CEO in his first quarter that he would cut the time needed for learning by 30% through better operations. That promise was specific and easy to track. It bought him support for the second, harder promise: a workforce that can carry out the strategy.
We also learned that concrete beats abstract. In one session, a leader heard a high-level pitch for “one governed view of people, work, capacity and impact” and was lukewarm. When she saw real numbers on screen (hours saved, full-time-equivalent capacity and the share of work tied to strategy) her reaction changed at once: “I’ve been trying to build something similar on my own team.” Lead with the numbers and the picture. Then the idea.
Three measurement models came up with leaders who are doing this well:
- Skills benchmarking
Quarterly self and manager assessments, rolled up into gap trends that guide build, buy or borrow decisions - Internal service provider
Measure the function like a service, with turnaround time, Net Promoter Score and a satisfaction score, applied the same way across every People function - Success defined up front
Set the measure before building, then run 30/60/90-day follow-ups with leading and lagging indicators so a weak programme is caught early
Actionable steps:
- Move your headline from activity to a tiered scorecard: efficiency and capacity first, behaviour change second, and business impact for flagship programmes third, with assumptions shown
- Roll impact up at three levels: one programme, a family of programmes such as leadership development, and the whole function against company goals
- Let the business own the business measure. A sales leader owns the sales number. L&D owns design quality, delivery and the evaluation method. Agree the baseline together before work begins
- Show your working. Separate modelled benefit from realised benefit, and say what is still uncertain. That is how CFOs read every other investment case
Tip: Research from Brandon Hall Group and Cognota found that organisations with a dedicated operations lead report 34% faster delivery, 28% lower costs and 39% higher stakeholder satisfaction. If you are asking for one hire, that is a good one to ask for.
7. Govern AI before it governs you
There is no typical starting point for AI. Leaders described three very different states:
- Chaos
A professional services firm left AI tools open across the company. It now faces client pushback over data, is paying about $5,000 a day in token overages, and is locking things down after the fact - Confidence
A credit union VP told us, “Everything we do is AI.” Her question is ‘what comes next’, not how to get control - Plateau
A consultant working with early adopters sees usage fading. People who were very engaged have stopped using the tools
The speed gains are real. A former CLO at two global tech firms once ran a 200-person learning team. Today her L&D team is four people. “For every one hour of course, it took 40 hours to design and develop it. Today we can do that in seconds or minutes.” Before backfilling any role, her team runs an “AI vulnerability audit” to see which tasks AI can take on first.
The safest path is a staged model: off, then AI-assisted with human review, then autonomous only once trust is earned, with an audit trail the whole way through. The organisations most likely to end up in the headlines are the ones that skip the stages and switch AI on everywhere from day one.
Leaders who are ahead share a few habits:
- They stage AI by person, not just by feature
One global consulting firm built three paths: general users, leaders and certified builders - They separate broad AI skills from deep work redesign
Prompting skills sit across the top. Rebuilding how a specific function works with AI runs deeper, and that work starts in the business - They teach people to rethink their work, not the tool
One team skipped prompt training and used short videos of peers already working differently - They propose the split with IT
One learning leader offered a deal: L&D leads AI fluency, IT owns the technology and governance. IT agreed and funded it
Actionable steps:
- Start with low-risk uses: summarising requests, flagging duplicate work, drafting status updates. Expand only once the data and review process prove reliable
- Run AI through an enterprise model your security team has already approved. It keeps data inside a known boundary and skips a new security review
- Keep a human check on all learner-facing AI content, and plan how reviewers will build the judgement to do that check well
- Run an AI task review before backfilling any role
Tip: Fix the process before you automate it. AI deployed on messy intake and unclear ownership just makes bad decisions faster.
8. Turn visibility into signals
The steps across both parts of this feature are not separate problems. They are stages in one loop, and each depends on the one before it. You can’t plan capacity for work nobody agreed on. You can’t measure impact from work nobody tracked. You can’t improve what was never measured.
We call it the operations framework, from visibility to signals. It works whether your function is central, federated or somewhere in between.
| Stage | The question to answer | What breaks without it | Owner |
| 1. Align | What work is happening, and does it connect to a business goal? | Demand arrives ad hoc, cut off from strategy | CLO / CHRO and directors |
| 2. Plan | Do we have the right capacity, budgeted and staffed, for what is coming? | The backlog fills with unranked work, and the busiest people absorb it | Director, L&D Ops |
| 3. Execute | Who owns this work, under what standard, and is it on time? | Work moves through fewer, more tired hands; vendor and AI risk goes unmanaged | Director, L&D Ops |
| 4. Measure | What business outcome did this produce? | Reporting stays stuck at completions and satisfaction scores | CLO / CHRO and CFO |
| 5. Optimise | Where are we on the maturity curve, and what is the honest next step? | Progress is a feeling, not a fact, and the same mistakes repeat | CLO / CHRO |
Use it as a diagnostic. A CLO fighting fires at Measure, while Align and Plan still run on instinct, is solving the wrong problem first.
The last step is the one most teams miss. Seeing everything is not the same as knowing what to do next. The leaders furthest along were not the ones with the biggest dashboards. They had found a way to turn a flood of activity into a short list of things worth acting on: a backlog quietly tipping into unranked work, a deadline about to slip, spend drifting outside its lane. Each signal goes to whoever is best placed to act on it. Sometimes that is a director who needs to move two tasks before Friday. More and more, for routine calls, it is a digital teammate the team has trusted with a narrow piece of the work.
Case study: A Fortune 50 financial services company
After a major reorganisation, this company’s learning team was supporting employees, field agents and business units through more than 1,600 active projects, with just two full-time project managers. Intake ran through spreadsheets, email and disconnected tools. Leadership had no consistent view of priorities, resourcing or links to business goals.
“We had no clear line of sight into volume, capacity, or business impact. It was impossible to prioritise or plan with confidence.”
Operations Lead
The team launched a LearnOps platform across the new organisation in 60 days. A standard intake portal replaced ad hoc requests. Real-time views showed who was working on what and how much capacity was left. Every project was mapped to one of four strategic business objectives. Dedicated operations roles, backed by expert support and a maturity assessment, drove adoption.
The team hit its full annual delivery volume by mid-year, on the same headcount. Business partners began submitting requests with context, timelines and expectations already attached, which cut weeks of back-and-forth. Leadership had the data to argue for funding, and the function went from being seen as a service desk to a data-driven partner.
A 90-day plan to get started
Don’t try to redesign everything at once. Pick one business-critical part of your portfolio and run a 90-day test.
- Days 1 to 30
Set a baseline. Measure incoming requests, the share of work with a priority, commitments, workload spread and business ownership. Agree priority criteria with business leaders - Days 31 to 60
Bring in one decision record and a regular cross-functional review. Every newly approved project must name a sponsor and an expected outcome. Document at least one real trade-off - Days 61 to 90
Compare against your baseline. Decide what to speed up, change, pause or stop.
The output is a decision log, not a dashboard. The goal is proof that visibility changes real decisions about people and money.
A quick self-check
If you answer no to any of these, you know where to start:
- Can you see total spend, work and capacity across the whole function in one place?
- Is every active project tied to a named company goal?
- Do you know what share of capacity goes to strategy versus reactive work?
- Is success defined before a programme is built?
- Would you know if another team launched its own learning programme?
- Is there a human check on AI-generated learning content?
- Could you show your CFO impact in one number, with the detail beneath it?
- Have you talked with your team about how their roles will change?
Don’t skip the people part
The hardest barrier we found was not price, ROI or IT. It was people. One CLO has believed in a new operating model for three years. What holds him back is whether his team can absorb the change. Many L&D professionals tie their sense of value to the manual, behind-the-scenes work a modern model takes off their plate. Two CLOs, separately, asked us the same question: can the people I have deliver this vision?
So, name the change openly. Involve the team in redesigning the work. Show examples of peers already working the new way. And keep the tools simple, so people don’t work around them or go back to old ways of working.
Earning through visibility
A seat at the table is not something you ask for. It is something you earn by showing the business, in its own language, what the function is working on, why it matters and what it produced.
That starts with visibility. Get it right and governance, prioritisation, capacity and measurement all become easier. Skip it and every new initiative, AI included, sits on an unstable base. A function that can’t see its own backlog today won’t be ready to hand routine decisions to a digital teammate tomorrow. It will just move the same blind spots faster.
The next frontier is human and digital teammates working the same operating loop side by side. That future isn’t a tool you buy. It’s a level of operational discipline you build, one decision at a time.
Ryan Austin is Founder and CEO of Cognota
The full report, Visibility to Lead. Signals to Execute, was produced with Brandon Hall Group

Good L&D thinking doesn’t stop here
Fresh ideas, expert insights, discover practical ideas you can use in your work, and explore perspectives to challenge your thinking
Get Training Journal’s Thursday newsletter straight to your inbox

