Artificial designs and features usually control production workflowsin the industry. The operator has the freedom to adapt designs toachieve the desired function; when the operator is satisfied, massproduction of the two objects is possible. The production workflowfor prosthetic restorations in dentistry is a fairly complicatedprocedure that requires several well-controlled processes, and eachunit is individually adapted to one unique situation. The aim of thefinal restoration is to replace damaged or missing soft and hard tissue,and to restore function, phonetics and aesthetics. The restoration hashigh material property requirements in order to withstand high forces,thermal changes, aging and humidity. If the fit of the reconstructionis insufficient there is a high probability for clinical failures rangingfrom inflammatory processes to reconstruction fractures. Thegrading of perfect, sufficient and insufficient fit is unknown althoughthe definition clinically acceptable fit has been used to describe andcontrol a reconstruction that is well seated and controllable by the clinician. Study I in this thesis focuses on the clearance (play) betweendifferent implant components in order to achieve a threshold value forhow accurate the production in dentistry needs to be. We found thatmetallic components on external hex connections have a clearanceof approximately 50 μm.Not only is every case individually designed and manufactured,but the receiving intraoral part also needs to be replicated into anextraoral part ahead of production, a procedure that has been possiblewith different impression materials. Subsequently, the production goes through a series of controlled compensations to fit the intraoralsituation. The conventional workflow starts by the selection of animpression tray, ranging from custom-made trays to plastic stocktrays. The ideal trays are rigid, thereby minimising flexure during theimpression taking. There are several types of impression materialswith different properties regarding setting time, volume changes andmechanical properties. The next step in the conventional workflowis the casting of the impression. There are various types of gypsumproducts utilised in dentistry, and they require different amounts ofwater. The differences depend on the shape and compactness of thecrystals. Type IV dental stone gypsum is often used in reconstructivedentistry with a typical setting expansion of 0.10%. For the partialdigital workflow the same volume changes can be seen for theconventional impression, the stone model production and the dieprocessing. In order to design the intended construction digitallyinstead of using wax, the model needs to be digitised in an extraoral scanner, also known as desktop scanner.The fully digital workflow consists of a direct digitisation of the oralcavity utilising intraoral scanning devices. All intraoral scanners havethe same goal, to digitise the size, shape and surface of a physicalobject into a geometrical virtual shape. This acquisition needs to berepeatable, reproducible and accurate. The IOS manufacturers tryto achieve these goals with different hardware and software setups.Study IV focuses on the acquisition accuracy of five different intraoralscanners for the digitisation of edentulous and dentated models. Theresults suggest that the devices had lower accuracy for the digitisationof the edentulous models when compared to the dentated model.Furthermore, Study IV presented observations suggesting that fullarchscans had lower accuracy when compared to shorter arch scanson both models. For the cross-arch measurements on the edentulousscans, the trueness values ranged from 6 μm to 193 μm, and, for the shorter arch measurements, the results ranged from 2 μm to 103 μm.For the dentated cast, the cross-arch trueness values ranged from6 μm to 150 μm, and, for the shorter arch measurements, the resultsranged from 4 μm to -56 μm.The digitised file is then utilised as a virtual model by a computeraideddesigner in order to virtually design the intended reconstruction.The designed file is then manufactured utilising computer-aidedmanufacturing, which can be performed either by a subtractivemachine (milling) or by additive systems (3D printing). Study II andStudy III explore the production tolerances for producing polymericand metallic objects from additive systems. Study III also containeda subtractive group. The results from these two studies suggest thatall tested additive systems for producing polymeric objects were, onaverage, 500 μm to