Booking product
Test Dates
A booking product for learner drivers looking for earlier DVSA test dates, with availability monitoring and a mobile-first interface.
Read the case studyVisit product (testdates.co.uk, opens in a new tab)
AI automation
Automate the repetitive path. Keep a person on the judgement.
Phigz connects models to the documents, inboxes and systems a team already uses, so repetitive preparation work moves faster without hiding the decision.
The problem
Operations teams lose hours moving the same facts between email, spreadsheets and line-of-business software. A generic chatbot does not sit inside that path.
Automation that writes straight into a finance, HR or customer system can also spread a bad extraction very quickly.
We automate the preparation and the routing. Where a result changes a record or reaches a customer, a person still confirms it.
What Phigz delivers
Workflow selection
We pick a repeated task with clear inputs and a visible output, not a vague ambition to “use AI everywhere”.
System connection
The automation reads from the source you name and writes only to the destination you approve.
A review surface
Exceptions, low-confidence results and high-impact actions are queued for a person.
A measured handoff
The team can see what ran, what was changed by a person, and what still needs a manual path.
Where it fits
Inbox and request triage
Classify incoming work, pull out the fields that matter, and route it to the right queue.
Document intake
Turn a recurring form or PDF into structured fields for the process that already checks them.
Internal updates
Draft status notes or CRM updates from activity the team already records, for someone to accept.
Delivery
01
Sit with the task
We watch how the work is done now, including the exceptions people handle quietly.
02
Draw the automatic stretch
Only the stable middle of the task is automated. The start and the commitment stay visible.
03
Connect and test
We run historical examples through the flow and compare them with what a person would have done.
04
Leave it with the team
The people who own the queue know how to pause it, correct it and see what it did.
Technical capability
Integrations
Connections are built against APIs and files you can give us access to. We do not assume a pre-built connector for every vendor.
Deterministic steps stay deterministic
Validation, formatting and routing rules remain ordinary code. The model is used where language or variation is the problem.
Failure behaviour
If a system or a model call fails, the item stays in a queue. It is not dropped and it is not silently written.
Related product work
Published case studies are product builds. Where a page is about AI, these projects show how Phigz ships a complete product. They are not labelled as AI systems.
Product design and platform
CheckMyRef
A platform for storing, managing and reusing employment references, with a brand built around clarity and trust.
Read the case studyVisit product (checkmyref.com, opens in a new tab)
Training platform
Bluetech Consulting
A training and career-support platform, designed around course discovery and a visual identity in professional blue.
Read the case studyVisit product (bluetechconsulting.com, opens in a new tab)
Questions
Will this replace the operations team?
No. It removes repetitive preparation. The people who understand the customer or the record still own the outcome.
Can you automate a process that changes every week?
A process without a stable core is a poor automation candidate. We will say that before building.
Is this the same as an agent?
Sometimes. A fixed workflow may be enough. An agent is worth it when the path branches and tools must be chosen along the way. We pick the simpler design that does the job.
Next step
Talk through this engagement.
Tell us the product, the current state, and the outcome you need. We will reply with a straight view of whether this is the right shape of work.