What’s holding you back?

Why this happens

We don’t know where to start.

AI touches everything, so everything looks like a candidate. Without a shared picture of where value is created and lost today, the organization defaults to the most visible tools or the loudest advocate.

What is probably causing it

No shared view of opportunities across leadership and teams
Ideas framed as tools and features instead of outcomes
Fear of choosing wrong, so nothing is chosen

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Discover

Find the opportunities worth your attention.

Identify the most valuable AI opportunities and the barriers in the way — and turn them into a roadmap connected to business outcomes.

Outcome: A prioritized, actionable roadmap connected to business outcomes.

Also consider: AI Inspire

Explore AI Discover

A valuable first experiment

A one-week opportunity map

Take a single value stream — from customer need to delivered result. Map where time, quality and money leak. Pick the one leak where AI plus a workflow change could plausibly help, and write it down as a hypothesis.

What we would measure

One prioritized hypothesis with a baseline metric, agreed by the people who do the work and the people who fund it.

How to frame it for leadership

“We propose starting with one value stream and one measurable hypothesis instead of an organization-wide AI program.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

Our employees aren’t using AI.

Access is not adoption. People use a new tool when it fits their work, when they feel safe doing so, and when they see leaders doing the same. Most rollouts deliver the access and skip the rest.

What is probably causing it

Training about AI rather than training in the work people actually do
Unclear rules — nobody knows what is allowed, so they wait
Leaders sponsoring the initiative without changing their own habits
Tools that don’t connect to the systems where the work lives

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Learn

From knowing about AI to working with it.

Develop practical AI capabilities for the roles that shape how work gets done.

Outcome: People confidently applying AI in their everyday work.

Also consider: AI Enable

Explore AI Learn

A valuable first experiment

One team, one recurring task, two weeks

Choose a team and one task they do every week. Redesign that task with them, with AI as a participant, agree what is allowed, and have the team’s leader do it too. Compare before and after.

What we would measure

Share of the task done the new way after two weeks; time and quality of the result; what the team says made it stick or not.

How to frame it for leadership

“Adoption follows redesigned work and visible leadership, not more licenses or another course.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We have too many AI tools.

Tools arrive bottom-up, team by team, each solving a local problem. Costs add up centrally; value stays local and invisible. Nobody has the mandate — or the data — to consolidate.

What is probably causing it

No shared platform or integration layer, so each team buys its own
Seat-based costs that grow regardless of use
No way to connect a subscription to an outcome
Governance that arrived after the tools did

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Enable

Foundations that make the next experiment cheaper.

Establish the technical, security and governance foundations for effective, safe AI-enabled work.

Outcome: Safe, scalable and effective AI-enabled work.

Also consider: AI Discover

Explore AI Enable

A valuable first experiment

A 30-day subscription review

Inventory every AI tool, who uses it, how often and for what. Pick the three biggest costs and pair each with a measurable use. Consolidate where two tools do one job. Agree the rule for the next purchase.

What we would measure

Tools per job, active use per seat, cost per active user — and a written decision rule for new tools.

How to frame it for leadership

“Fewer, connected tools with measurable use will cost less and deliver more than the current sprawl.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

Our development teams are faster, but delivery isn’t.

AI accelerated the step it was pointed at — writing code. The system still delivers at the pace of its slowest step: review, testing, release, decisions. Local speed became a queue somewhere else.

What is probably causing it

Bottlenecks in review, verification and release that were tolerable at the old pace
Decisions and approvals that wait for the same meetings as before
Quality concerns about AI-generated code that slow review further
Pipelines designed for human-paced change

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Experiment

Evidence before investment.

Validate AI opportunities through hands-on experimentation in real workflows, with real teams, measured against real outcomes.

Outcome: Validated solutions and evidence of measurable improvements.

Also consider: AI Enable

Explore AI Experiment

A valuable first experiment

Move the bottleneck, then remove it

Map the flow from idea to production for one team and find where work waits. Redesign that step — often review or verification — with AI assistance and clearer quality rules, then measure end-to-end lead time instead of coding speed.

What we would measure

Lead time from commit to production, work-in-progress at each stage, and change failure rate — before and after.

How to frame it for leadership

“Faster coding only pays when the whole delivery flow speeds up. We propose measuring lead time, not output.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We need better AI governance.

Policies were written for a world without AI, so they say no by default. Teams either stop or route around them with personal tools. Governance becomes the bottleneck — and the scapegoat.

What is probably causing it

Security, legal and teams designing rules separately
No data classification people can actually apply
Uniform caution instead of risk proportionate to the use case
No owner for quality of AI-assisted work

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Enable

Foundations that make the next experiment cheaper.

Establish the technical, security and governance foundations for effective, safe AI-enabled work.

Outcome: Safe, scalable and effective AI-enabled work.

Also consider: AI Scale

Explore AI Enable

A valuable first experiment

Guardrails for one use case

Take one real, valuable use case that is currently blocked. Put security, legal and the team in one room and design the minimum guardrails that make it safe. Ship the guardrails and the use case together.

What we would measure

Time from request to approved use, number of shadow tools in that area, and incidents — plus whether the pattern can be reused for the next case.

How to frame it for leadership

“Governance designed with the teams enables safe adoption; governance designed around them creates shadow AI.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We struggle to measure AI ROI.

Most AI initiatives measure activity — licenses, prompts, hours saved on a survey. Business outcomes live elsewhere and were never connected. Without a baseline, there is nothing to compare against.

What is probably causing it

Initiatives started without an outcome hypothesis
No baseline measured before the change
Productivity gains that never reach a customer or a budget line
Costs tracked centrally while value stays local

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Discover

Find the opportunities worth your attention.

Identify the most valuable AI opportunities and the barriers in the way — and turn them into a roadmap connected to business outcomes.

Outcome: A prioritized, actionable roadmap connected to business outcomes.

Also consider: AI Scale

Explore AI Discover

A valuable first experiment

Instrument one running initiative

Choose one AI initiative already under way. Write the outcome it is supposed to move, measure the baseline this week, and define the business metric it should eventually affect. Review in a month.

What we would measure

One outcome metric with a baseline and a trend — and a reusable template for the next initiative.

How to frame it for leadership

“We can show return when initiatives are framed as outcome hypotheses with baselines, not as tool rollouts.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We have successful pilots but cannot scale.

A pilot proves something is possible under special conditions: a champion, slack time, forgiving stakeholders. Scaling requires the organization — leadership, operating model, funding, governance — to change so the result repeats without those conditions.

What is probably causing it

Success tied to individuals rather than to a documented way of working
Other teams lacking the tooling, permissions or skills the pilot team had
No funding or portfolio mechanism to move from pilot to standard
Leaders sponsoring pilots but not changing how they run the business

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Scale

From one team’s success to the organization’s capability.

Turn successful experiments into organizational capabilities: leadership, operating model, culture and measurement.

Outcome: Sustainable improvements across teams and the organization.

Also consider: AI Partner

Explore AI Scale

A valuable first experiment

Second team, same outcome

Take the pilot to one more team with a different context. Document what had to change — tools, rules, skills, leadership behaviours — for the result to repeat. That list is your scaling plan.

What we would measure

Whether the second team reaches a comparable outcome, how long it took, and which enabling changes were required.

How to frame it for leadership

“Scaling is an organizational design problem, not a communication problem. We propose learning what it takes with a second team before a broad rollout.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We want to redesign how our teams work.

AI changes what a team can do, which changes what the team should look like: roles, hand-offs, rhythms and decisions. Treating AI as a plug-in to today’s process leaves most of that potential unused.

What is probably causing it

Processes designed around hand-offs that AI makes unnecessary
Roles defined by activity rather than by outcome
Decision rhythms unchanged while analysis got faster
No safe space for a team to try a different design

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Experiment

Evidence before investment.

Validate AI opportunities through hands-on experimentation in real workflows, with real teams, measured against real outcomes.

Outcome: Validated solutions and evidence of measurable improvements.

Also consider: AI Learn

Explore AI Experiment

A valuable first experiment

Design a workflow from scratch, with AI in the room

With one team, take one workflow and ask: if we designed this today, with capable AI as a participant, what would it look like? Prototype the new design for a sprint and keep what works.

What we would measure

Cycle time, number of hand-offs, team satisfaction and quality of the result — compared with the old design.

How to frame it for leadership

“Redesigning team workflows is where productivity gains become business results. We propose one team as a learning environment.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We want to explore AI agents.

Agents are most valuable where work is repetitive, rule-rich and spread across systems — and most dangerous where nobody has thought about permissions, quality and oversight. Exploring them is as much about the work and the guardrails as the technology.

What is probably causing it

Curiosity about agents without a specific bottleneck for them to remove
Systems that agents cannot safely reach
Unclear accountability for what an agent does
Skills gap in designing agentic workflows

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Experiment

Evidence before investment.

Validate AI opportunities through hands-on experimentation in real workflows, with real teams, measured against real outcomes.

Outcome: Validated solutions and evidence of measurable improvements.

Also consider: AI Enable

Explore AI Experiment

A valuable first experiment

One agent, one bottleneck, human in the loop

Pick a repetitive, well-defined task that crosses systems and frustrates a team. Build an agentic workflow for it with explicit permissions and a human checkpoint. Run it on real work for a few weeks.

What we would measure

Tasks completed by the agent versus escalated, quality of results, time saved for the team and what the guardrails caught.

How to frame it for leadership

“A contained agent experiment tells us where agents pay off and what governance they need — before any broad investment.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

Why this happens

We need to improve software quality.

AI increases the volume and speed of change. If quality practices — specification, testing, review — don’t evolve at the same pace, defects rise and trust in AI-assisted work falls. Quality becomes the reason not to adopt.

What is probably causing it

Testing and review designed for human-paced, human-written change
Vague specifications that AI amplifies into confident, wrong code
No agreed definition of quality for AI-assisted work
Pipelines without the checks that fast change requires

Value Model drivers involved

The first dimension is usually where the lever is. The others decide whether the lever holds.

An offering that could help

AI Enable

Foundations that make the next experiment cheaper.

Establish the technical, security and governance foundations for effective, safe AI-enabled work.

Outcome: Safe, scalable and effective AI-enabled work.

Also consider: AI Experiment

Explore AI Enable

A valuable first experiment

Quality-first AI development on one service

For one service or component, redesign the development workflow so AI helps write specifications and tests before it writes code, and review focuses on intent and risk. Strengthen the pipeline to match.

What we would measure

Change failure rate, defects found before and after release, review time and test coverage — compared with the previous quarter.

How to frame it for leadership

“Speed without quality is a liability. We propose making quality the first thing AI improves in the delivery flow.”

This page remembers your selection in the link — send it to whoever needs to see it.

Let’s explore this together

What’s holding you back?

We don’t know where to start.

What is probably causing it

AI Discover

A one-week opportunity map

Our employees aren’t using AI.

What is probably causing it

AI Learn

One team, one recurring task, two weeks

We have too many AI tools.

What is probably causing it

AI Enable

A 30-day subscription review

Our development teams are faster, but delivery isn’t.

What is probably causing it

AI Experiment

Move the bottleneck, then remove it

We need better AI governance.

What is probably causing it

AI Enable

Guardrails for one use case

We struggle to measure AI ROI.

What is probably causing it

AI Discover

Instrument one running initiative

We have successful pilots but cannot scale.

What is probably causing it

AI Scale

Second team, same outcome

We want to redesign how our teams work.

What is probably causing it

AI Experiment

Design a workflow from scratch, with AI in the room

We want to explore AI agents.

What is probably causing it

AI Experiment

One agent, one bottleneck, human in the loop

We need to improve software quality.

What is probably causing it

AI Enable

Quality-first AI development on one service

Not on the list?