Reverse logistics for hospital pharmacy

Every return,
in its place.

Assisted identification, guided sorting and traceability for returned medicines.

PharmacOps aims to delegate repetitive tasks to software and focus attention on exceptions. From evidence to destination, with every step recorded.

See the workflow in action Interactive prototype
T01DIGITAL TWIN
16/20 OCCUPIED POSITIONS · 5 × 4
DESTINATION · ITEMS
SuitableReviewDiscard

Group by destination and prepare for restocking.

Watch it fill automatically

Recognise the item.
Determine its route.

Handling a return means identifying the item, deciding where it belongs and recording what happened. PharmacOps brings these tasks into one workflow that moves clear cases forward and flags those needing attention.

Capture

Gather the sample evidence.

Identify

Suggest a reference and flag uncertainty.

Route

Assign a compatible position.

Trace

Keep the decision and every movement on record.

The workflow moves on.
Exceptions stand out.

Continue. Review. Separate.

PHARMACOPSWORKFLOW STATUS
DECISION+ REASON
WAITINGScroll to follow the workflow
CONTINUE

Clear cases keep moving.

Identify, assign a destination and record the route: the software handles repetitive tasks. In the planned first phase, physical placement remains manual.

Suitable for the workflow → destination and traceability
REVIEW

Attention only where needed.

Insufficient evidence or a discrepancy triggers a review. The exception stays located while subsequent samples continue. Resolving it leaves a decision and a reason that can inform future improvements.

Insufficient evidence → targeted review
SEPARATE

Items that cannot return are separated.

Expired, opened, damaged or unapproved items leave the restocking workflow. The cause is recorded and the issue remains traceable.

Not restockable → separation and recorded cause

Start with software.
Scale into physical automation.

A single model connects decisions, positions and movements. Physical automation will be introduced in modules once the process and its rules have been validated.

Organise the workflow.

The demo already explores sorting, destinations, trays and traceability. Physical movement in the planned MVP will be manual.

Visual demo and automatic mode
Available
Reservations, states and logs
Working with examples
Live capture, catalogue and identification
Awaiting integration

Automate movement.

Input, sorting and physical execution can be added as modules using the same position model. Each module will require its own testing.

INPUTCONTROLT01 / 5 × 4Phase 2 concept diagram
Explore the modular architecture Capture, identification, destinations and traceability
CAPABILITY / 01

Capture and evidence

Collects the image and preserves the information observable in it.

Planned input
Planned individual capture
Planned output
Observable evidence
Current status
In development

Product outline. The status shows what has been tested in the demonstration and what requires integration.

Learning from exceptions An opportunity for future improvement

Every resolved exception leaves a useful signal.

The final reference, reason and any correction help identify where the process fails and guide improvements. Professional judgement comes in when a case needs review.

  1. Proposal
  2. Exception
  3. Resolution and reason
  4. Structured feedback
  5. Evaluation and improvement

The demo records these signals. Model evaluation and training remain pending; there is no real-time machine learning.

A hackathon demo.
A project awaiting validation.

Born from LabORA Challenge 6 on medicines returned to hospital pharmacy.

We aim to assess the workflow with professionals and prepare future tests using authorised samples.

Let's talk about what we can build together.

PharmacOps welcomes collaboration to validate the process and develop its next phases.

[email protected]

Hospitals and hospital pharmacy

We are looking for hospitals, pharmacy services and healthcare professionals to help validate PharmacOps against real medicine return, identification and sorting processes.

We want to understand today's workflows first-hand, identify real needs and explore future pilot studies.

Robotics, automation and technology

We are looking for engineers, robotics companies, integrators, technology providers and entrepreneurs who want to help develop the physical side of PharmacOps.

From computer vision and sensors to sorting and automation systems, we want to build an architecture ready to scale into real environments.

Investors

PharmacOps is at an early stage and open to discussions with investors interested in healthtech, medtech, robotics, deeptech and automation.

We are looking for partners who share our vision of transforming a largely manual pharmaceutical process into scalable technological infrastructure.

Other collaborators

Do you work in pharmaceuticals, healthcare, technology or research and think we could build something together?

We welcome universities, technology centres, researchers, suppliers and others who can contribute knowledge, technology or new collaboration opportunities.

Tell us what's on your mind.

Choose your profile and send your message here. You can also email [email protected].

By submitting, you share these details with PharmacOps through Cloudflare and Idriss's email. See Privacy. Do not send patient data or clinical documentation.

The essentials,
clearly explained.

Does the MVP already include a robot?

No. The planned first phase combines a mobile camera, software and physical trays with human placement. Connected capture and mechanical automation are future integrations.

What does the demo demonstrate?

A sample's journey to its destination: identification, grouping by zone and traceability in T01. Follow it step by step or use automatic mode with speed and pause controls.

Does it learn automatically from each review?

The demo records the responsible person, decision, reason, conflicting evidence and corrections. This information could support future evaluations; today there is no automatic training, learning pipeline or measured accuracy improvement.

What happens when a sample cannot be identified?

The case remains located and awaiting review while subsequent samples can continue. Intervention focuses on that exception; insufficient evidence does not become a confirmed identity.

Does a green result mean the medicine can be dispensed?

No. Green indicates suitability to continue in the logistics workflow; dispensing requires its own professional procedure.

Is it deployed or endorsed by Hospital Clínic?

It is not presented as a deployment or an endorsed solution. Hospital Clínic promotes the challenge that inspired the project.