The ones that come up on almost every first call, answered here so the call can be about your system instead.
engineering / the practice
Senior engineers, named in the statement of work, and the delivery lead writes code too. There is no arrangement where a senior team wins the work and a junior team delivers it. If a name on the proposal changes, you hear it from us before it happens and you see the handover.
Our engineers use AI tooling in the same way they use a compiler and a debugger, and every line is reviewed, tested and owned by the person who put their name on the pull request. The phrase we use internally is that we are AI augmented, not AI washed. The difference shows up in the review queue.
The team is based in India. Standups and live calls sit in your morning if you are on the United States east coast, and inside your working day entirely if you are in Australia or the Gulf. The rest of the day runs on written handoffs, so work continues after your team logs off and there is a note waiting when they log back on. It is an advantage when the handoff discipline is real and a liability when it is not, so we treat the handoff note as a deliverable.
Yes, in full. IP assignment is in the contract, the repository is yours from the first commit, and we retain no licence to reuse anything built for you. Nothing we deliver depends on a Thinkscoop hosted service.
During an engagement you get runbooks for the failure modes that specific system has, alerts routed to people who can act, and our participation in incident review. Around the clock rotation is not included by default because pretending otherwise is how a supplier misses a page. If you need it, it is a scoped line on an Embedded AI Pod.
The major hosted providers and open weight models running in your own environment. The architecture treats the provider as a configuration decision behind one interface, so the answer to this question should stop mattering to you shortly after go live. A version named on a website is out of date within the quarter, so these pages name the provider and not the version. Where an older case study elsewhere on the site does name one, read it as a snapshot of what that engagement ran on, not as a commitment about what yours will.
Yes. NDAs, your access control, your tenancy, and data flow documentation written for your reviewer rather than for us. Tell us in the first conversation that a review is coming and it gets scoped into the plan. On certification: the SOC 2 Type I audit is in progress and not yet certified. We would rather you read that here than discover it in week ten.
Then that is the answer you get, in the scoping call or in the written brief two days later. It has cost us engagements. It has also produced most of the second projects.
engineering / money
Because the alternative is that an engineering leader with a real project has to sit through a discovery call to find out whether the number starts with a two or a two hundred. Both sibling practices at Thinkscoop publish theirs. Withholding it filters for patience rather than for fit.
The number of external systems that have to be integrated, whether the build has to run inside your own tenancy from day one, how much of the process exists only in people's heads at the start, and the size of the compliance surface. Not how large your company is.
Model provider and infrastructure costs, which you pay directly on your own accounts, so there is no margin on your usage and no incentive for us to be relaxed about your token spend. Third party licences are also yours and named in the scope before signature.
The sprint is fixed fee in instalments against milestones. The integration pod and the workflow build are billed monthly against a fixed capacity, with the workflow build carrying a stopping point at the end of design. The embedded pod is monthly in advance, three month minimum, then rolling with 30 days notice. USD throughout.
All four ranges are on the engagements page.
Ask it directly. If it is a good one it usually ends up on this page.