Dako's AI Builder Notes
⌘K
中EN
System
Color theme
Changsha HubOffline
大可 Dako
大可 Dako大可 DakoAI Builder · Agent PractitionerTurning AI into working software. Practical notes on agent engineering, vibe coding, and commercial delivery.
Platform links coming soon
MVD keeps the depth and narrows the slice to where value can appearReal dataUse data the customer actually works withA sample working is not a field resultDeep narrow sliceOne workflow, end to endDo not make MVD a crippled planFixed deadlineVerify in weeks, not half a yearLong timelines invite scope growthThe judge is continued use in the real workflow, not a convincing demo
Real data, narrow the scope without reducing depth, and set a hard deadline.

Minimum Viable Deployment

MVD uses the smallest serious engineering effort to verify that value appears in a real customer environment.

An MVP asks whether a product should exist in a market. An MVD asks whether a solution can create value for this customer in this environment. The difference is the judge.

After reading, you should be able to answer:

  • How is MVD different from MVP, and why does "make it run first" matter?
  • Why is the MVD validation window counted in weeks rather than months?
  • How can early usage open the door to wider adoption?

Rule one: real data, no exceptions

Validating on anonymized samples or invented data is the first brick in a proof-of-concept grave. Real data contains the details that break plans: undocumented field meanings, missing values, and contradictory "correct" answers. The customer must bring the data it actually uses. A solution that works on fake data meets the real price on launch day.

Rule two: narrow the scope, not the depth

MVD is not a half-built version of a large plan. Keep the engineering depth and reduce the surface area: do not cover "intelligent support for the whole company"; cover returns and exchanges end to end. The slice should be small enough to build and deep enough for a business team to see the value. The legal AI company Harvey expanded along exactly that path: UK law firm Ashurst ran a trial across the whole firm, and only after it held up announced a global partnership ashurst-harvey. Prove one bounded thing credibly, then spread sideways.

Rule three: set a hard deadline

MVD should be measured in weeks, not months. Palantir's field sprint — the industry calls it a bootcamp — is the extreme form: the customer brings real data, the FDE team works on site, and together they produce a prototype that runs and can be clicked through live within days. The point of a fixed window is not speed. It forces honest trade-offs. Anything that cannot show value within the window is probably not the core value. A six-month "minimum validation" almost always grows back into a project that wants everything.

References

  1. Ashurst launches global Harvey partnership following extensive firmwide trial