Contact Centre Admin — one console out of ten tools
A market-leading on-prem contact centre had grown its administration across separate product workstreams until it behaved like ten products rather than one. A competitor was selling against exactly that. The fix started with how design worked, not with the screens.
The consolidation answered in structure, not styling. Above: administration regrouped into eight areas an admin can hold in their head — infrastructure, call settings, email and chat, users, organisation, bulk import, desktop settings, features — rather than a menu inherited from whichever team happened to build each part. Below: people and permissions in one table, every role a badge on the person, so who-can-do-what is read rather than reconstructed across separate tools.
designers inherited — embedded in the product triads, but with no real say in what got built.
workstream used to pilot the research model, before asking for anything wider.
workstreams the model reached once the pilot had proved itself.
disjointed tools the agent-side experience was spread across before consolidation.
-
Rebuild the practice before the product
The first diagnosis was not about screens. Design sat inside the product triads without influence over what got built, so any redesign would have been overwritten by the next roadmap. Changing how design worked came first.
-
Win an executive sponsor
In an engineering-led organisation, a design argument travels only as far as the person willing to carry it. Securing a sponsor at the executive level turned research from a request into an expectation.
-
Pilot the research model on one workstream
Rather than proposing a programme, we ran it once — on a single workstream, small enough to fund out of goodwill. The pilot was the proposal.
-
Scale the model to four
With one workstream's results in hand, the same model extended across four. Each adoption made the next easier to ask for, because the ask was now a precedent rather than an idea.
-
Introduce an end-to-end journey lens
The org was organised by product area, so nobody was looking at the administrator's journey across it. Mapping the whole path made the fragmentation impossible to argue with — and gave the consolidation a shape to aim at.
-
Collapse the tools into one interface
The disjointed agent-side tooling was folded into a single unified interface with a lighter deployment, so the thing being sold matched the thing being demonstrated.
- Administration had grown independently inside each product workstream — nobody owned the whole of it.
- A competitor was actively exploiting the gap, and sales was citing experience as a reason for lost deals.
- Design was embedded in the triads in name only: present at the table, absent from the decision.
- The platform's analyst leadership position was under real scrutiny, so the work carried commercial weight from day one.
Disjointed tools collapsed into one interface, with a lighter deployment behind it. The platform held its market lead under sustained competitive and analyst pressure — and sales stopped citing the experience as a reason for lost deals. The research model outlived the project: it became how the four workstreams worked, not a thing that happened once.
Fix the practice before the product
Redesigning the admin first would have produced a better set of screens inside the same broken decision structure. Winning influence over what got built was the precondition, not the reward.
Pilot small, then let the evidence ask
One workstream, run properly, made the case that no deck could. Scaling to four was a series of small, already-answered requests rather than one large unfunded one.
Argue from the journey, not the feature list
In an org sliced by product area, the end-to-end journey was the one view nobody owned — and the only one in which ten tools obviously read as a single broken experience.