Outcomes
No engagement outlives its gate.
Consulting engagements renew. Platforms get extended. Both can run for years without anyone asking whether the original goal was met. Every engagement we take has an outcome, a number and an end date agreed before it starts. At day 90 the evidence either supports going on or it does not, and that decision belongs to your executive committee rather than to us.
The boundary
The terms flex. The discipline does not.
How an engagement starts, what you hold at the end, and how it is contracted all depend on where you are: sometimes a day rate, sometimes a fixed outcome, sometimes a mix. What never changes is that there is an outcome you name, a number you set, and an end date agreed before the work begins. Inside those boundaries the structure follows what the result is worth to you.
The choice
Most value is lost choosing the wrong problem.
The first work is not delivery. It is identifying which problem, solved, would be worth the most to your business this year. The solution is chosen second. That decision is worth more than the execution that follows, and it is where most programmes are lost before they begin.
What you hold at day 90
A boutique leaves you a working system. We leave you the decision.
A working system is where a boutique stops, and it is where the outcome usually stalls. At day 90 you hold more than software: a measured result against the baseline you started from, a business case with the cost of ownership, a team of your own people trained to run it, and the system itself in production. Not every one on every engagement. The invariant is that the evidence an executive committee needs to decide exists, and you own it.
The role of technology
We do not sell AI. We use it where it pays.
AI and automation are instruments. We choose them when they earn a return and reject them when they do not. A recommendation to buy rather than build, or to do less than you planned, is a legitimate result of the work and one we have delivered more than once. We can make it because Bartholdi Labs holds no software licence revenue and because our ventures arm gives us a route to what already exists. Total cost of ownership and return decide the question. Where owning the intellectual property is not the point, we say so.
Who does the work
Operators who have run the function, inside the team.
The people who have carried the responsibility themselves sit inside the delivery team while the work happens, not at a steering committee reviewing it. Someone who has already run the function being fixed. Behind them, 10+ Bartholdi Certified senior mentors and a network of 90+ experts. This is the difference a boutique cannot staff.
Delivered outcomes
Customer service
Health and fitness · More than 180 clubs
No unified member data. High churn. Reactive support across the estate with no AI.
An AI member care agent resolving 70 percent of queries autonomously, plus a churn predictor flagging members at risk 45 days early.
- Support cost down 32 percent
- Member retention up 18 percent
12 weeks to production
Supply chain
Renewable energy · European infrastructure group
Multi supplier procurement with no demand forecasting. Margin erosion and supplier risk blind spots.
An AI procurement optimiser, a demand forecasting model and supplier risk scoring, deployed in one sprint.
- Procurement cost down 15 percent
- Forecast accuracy up 40 percent
6 week pilot to production
Commercial
Premium beauty · 3,500 sites, 15 countries
Sites losing revenue to empty capacity. No cross sell engine. Multi brand data siloed.
Dynamic capacity pricing, an AI clienteling engine and a basket size optimiser, deployed across all brands.
- Revenue per seat up 23 percent
- No show rate down 28 percent
3 use cases in 90 days
Client identities withheld under mutual confidentiality. Executive references provided on request. Every indicator below was baselined before the work started.