The challenge: interactive demo (2 min)

Evidence

Are your teams really learning? We measured it.

A learning method validated by 5 industrial PhD theses, evaluated against Microsoft Dynamics 365 Guides, and results measured during our deployments. The goal is not to make people act like robots: it is for everyone to understand, remember and make the skill their own.

Our approach

First, understand how a hands-on skill is learnt. Only then, the tool.

Before designing our tool, we studied how people learn to do with their hands what they cannot yet do.

  1. A learning method from our research work

    With research laboratories, we studied how people learn a task they don’t know how to do, in a constrained environment. The aim is not to make people act like robots: it is for everyone to understand, remember and own the skill. This method is validated by 5 industrial PhD theses.

  2. A creation method within every expert’s reach

    The real challenge is creation: experts don’t have time, and not all of them are comfortable with digital tools. They create their own tutorials, very quickly, with the help of the built-in AI, and these tutorials apply the learning method without them needing to know it.

  3. The tool we wish we’d had

    As former industry professionals, we know every second counts: the expert’s time must be kept as short as possible. We also know the value of trade or defence secrets: everything works offline, without compromise, artificial intelligence included, right down to translating a tutorial in a few seconds.

The result: tutorials created in a few minutes, that really make people learn.

Measured results

What our deployments have measured.

  • < 6 monthsReturn on investmentObserved in our deployments
  • After 1 day8 trainers ready to work after one dayMeasured during our deployments
  • 1 minTo create an instructionBy the expert, on the real equipment, with no writing or editing (measured in our deployments)
  • × 2Train twice as fast, autonomouslyMeasured in our deployments, at a tier-one automotive supplier
  • ÷ 5Trainers and mentors tied up5 times fewer, measured in our deployments
  • +90%RetentionAccording to our method, validated by 5 industrial PhD theses

Scientific evidence

Evaluated against Microsoft Dynamics 365 Guides.

Our method was compared with Guides in a study published at IEEE ISMAR 2022, the leading international conference on augmented reality. It stems from work carried out with Institut Mines-Télécom in a boiler manufacturing plant. SHIFT89 was also incubated by IMT Starter, the Institut Mines-Télécom incubator.

ATOFIS, an AR Training System for Manual Assembly: A Full Comparative Evaluation against Guides (new window) — IEEE ISMAR, October 2022.

With 16 participants, compared with Guides: a tutorial created 1.72 times faster, no assembly errors and 13% faster learning, with better usability and mental workload scores.

The publication on the Institut Polytechnique de Paris portal (new window) · All our publications

  • 5 industrial PhD theses
  • 5 scientific publications (IEEE, Springer, ComSIS)
  • Contributions to ETSI and ISO/IEC standards
  • +90% retention

Publications

Our scientific contributions.

Our research, carried out with Institut Mines-Télécom, has been published at international scientific conferences and in an international scientific journal, and has contributed to the standardisation of augmented reality.

Scientific conferences

  1. ATOFIS, an AR Training System for Manual Assembly: A Full Comparative Evaluation against Guides (new window)T. Lavric, E. Bricard, M. Preda, T. Zaharia. IEEE International Symposium on Mixed and Augmented Reality (ISMAR), Singapore, 2022, pp. 558-567 · DOI 10.1109/ISMAR55827.2022.00072
  2. An Industry-Adapted AR Training Method for Manual Assembly Operations (new window)T. Lavric, E. Bricard, M. Preda, T. Zaharia. HCI International 2021, Late Breaking Papers, LNCS 13095, Springer, 2021 · DOI 10.1007/978-3-030-90963-5_22
  3. Exploring Low-Cost Visual Assets for Conveying Assembly Instructions in AR (new window)T. Lavric, E. Bricard, M. Preda, T. Zaharia. IEEE INISTA, Kocaeli (Turkey), 2021, pp. 1-6 · DOI 10.1109/INISTA52262.2021.9548570
  4. An AR Work Instructions Authoring Tool for Human-Operated Industrial Assembly Lines (new window)T. Lavric, E. Bricard, M. Preda, T. Zaharia. IEEE AIVR, Utrecht (Netherlands), 2020, pp. 174-183 · DOI 10.1109/AIVR50618.2020.00037

Scientific journal

  1. A Low-Cost AR Training System for Manual Assembly Operations (new window)T. Lavric, E. Bricard, M. Preda, T. Zaharia. Computer Science and Information Systems (ComSIS), vol. 20, no. 1, 2022, pp. 1047-1073 · DOI 10.2298/CSIS211123013L

Standardisation

  1. Industrial use cases for AR applications and services (new window)ETSI GR ARF 002 V1.1.1, July 2019 (PDF)
  2. AR standards landscape (new window)ETSI GR ARF 001 V1.1.1, April 2019 (PDF)
  3. AR framework architecture (new window)ETSI GS ARF 003 V1.1.1, March 2020 (PDF)
  4. Information technology — Multimedia application format (MPEG-A) — Part 13: Augmented reality application format (new window)ISO/IEC 23000-13:2017

Deployment

Up and running the next day.

One day of training

Your trainers create their first tutorial that same afternoon, then, from the next day, on their own, as often as is useful. In our deployments: 8 trainers up and running after one day.

Your procedures as the basis

Your existing procedures are not lost: the expert enriches them with their hands-on skill and key points of attention, on the equipment.

Living tutorials

Editing a tutorial is even faster than creating it. It is updated in a few minutes.

To go further

Compare, quantify, check.

Your best expert’s know-how deserves to be preserved.

A 30-minute to 2-hour demo, on one or more of your real cases, in your environment, with your teams. Just agree on a date.

Request a demoTake the challenge (2 min)

or write to us