Case studyClean energyB2B SaaSDesktop web app

Gridcog is simulation software for planning energy projects: solar, batteries, wind, EV charging and the markets they trade in. I've worked with the product team on the planning app, from the first version under the name Gridcognition to today's project designer and library.

gridcog.com ↗
Client
Gridcog
Energy modelling software, London and Perth
My role
Product designer
Product strategy, UX and UI
Scope
Planning app
Project designer, library and chat
In use
18,000+ projects
Modelled across 40+ markets, per Gridcog
Gridcog's project designer showing three scenarios, each with a physical, commercial and optimisation model
Fig. 1The project designer

The problem

A warehouse roof covered in solar panels, with two battery cabinets and an electric van charging in the car park
One small site: solar, two batteries and an EV charger to model together.

An energy project used to be one asset selling into one market. Now it is solar with a battery beside it, charging from the grid when prices are low, selling into several markets at once, and splitting the money between an owner and an optimiser.

Working out whether that stacks up means modelling the physical site, the commercial deal and the operating strategy together, across dozens of scenarios. Many teams still do it in spreadsheets that only their author can follow.

Gridcog replaces the spreadsheet. Its users are experts who need every parameter, and the answer still has to make sense to the people approving the spend.

A solar farm on red earth with a row of battery storage containers and a substation behind it

A model has hundreds of inputs. The screen should show the few that matter right now.

What I designed

  1. 01

    Projects

    Where work starts

    Every project a team is planning, as cards with their status and comments.

  2. 02

    Project designer

    Where the model is built

    Scenarios, each with a physical, a commercial and an optimisation model.

  3. 03

    Library

    Where assumptions live

    Tariffs, market prices, load profiles and costs, shared across projects.

  4. 04

    Chat

    Where questions get answered

    Team comments and answers from the documentation, beside the work.

The decisions that mattered

01 / 05

Every scenario is three models

A scenario is a row of three boxes: the physical model (what is on site), the commercial model (who pays whom) and the optimisation model (how it is run). Scenarios stack down the page, so a baseline sits directly above the options being tested against it.

Instead of
One long form for each project
I chose
Three models per scenario, stacked to compare
Three scenarios stacked, each a row of physical, commercial and optimisation model boxes
Fig. 2A baseline and two solar-and-battery scenarios, one above the other
02 / 05

The site is drawn as it is wired

The physical model is a diagram: the site, its meter, and the battery and solar hanging off it. Adding an asset means choosing from a grid of types, from a simple load to a wind turbine. It reads the way an engineer would sketch it.

Instead of
A table of assets
I chose
A diagram of the connection
The physical model as a tree: the site on a map, its meter, then a battery and solar
Fig. 3A site, its meter, and the battery and solar connected to it
03 / 05

Money flows left to right

The commercial model shows each participant as a box, with income arriving on the left and costs leaving on the right. Splitting revenue between an asset owner and an optimiser is a percentage on the arrow. Colour separates wholesale energy, ancillary services and hardware.

Instead of
A form of contract terms
I chose
A flow you can read like a ledger
Income on the left and costs on the right of the asset owner and optimiser, with the split shown as percentages
Fig. 4Income and costs for the asset owner and the optimiser
04 / 05

Assumptions live in a library

Tariffs, market prices, load profiles and cost assumptions are kept in a shared library and reused across projects. When a modeller changes one inside a project, the save button asks whether the change applies to this scenario only.

Instead of
Assumptions retyped in every project
I chose
One library, with local overrides
Four library tiles: load and asset profiles, electricity supply arrangements, wholesale markets and project costing assumptions
Fig. 5The library: reusable assumptions, by kind
05 / 05

Help arrives inside the project

Each project has a chat panel for the team. Ask it a modelling question and the answer comes from Gridcog's documentation, as steps that name the exact settings to change. If the answer falls short, one click sends the question to Gridcog's industry team.

Instead of
A help centre in another tab
I chose
Answers in the project, from the docs
A question about DC-coupled hybrid projects answered from the docs as numbered steps, with a button to send it to the industry team
Fig. 6A modelling question answered from the documentation
Gridcog's results view on a laptop, with a returns chart and a map of sites around Melbourne
A person working on a Gridcog site diagram on a laptop
Gridcog's sign-in screen on a pink and purple gradient
Gridcog's project grid on a laptop in a dark room

01 / 04

More of the product

Six more screens from the planning app and the library.

A grid of project cards with images, status and comment counts
ProjectsEvery project, with its status and comments.
A dialog for choosing an asset type, including battery, EV supply equipment, flexible load, genset and wind
Asset typesFrom a simple load to wind and EV charging.
A battery set-up form with sizing, duration, control mode and cost fields
Battery set-upSizing first, technical detail folded away.
Optimisation settings with participant choices and an uncertainty mode switch
OptimisationWho the model optimises for, and how certain it is.
Wholesale market categories: energy, ancillary services, capacity and balancing markets
MarketsEnergy, ancillary, capacity and balancing.
A day-ahead price chart for a wholesale market, fed from a live profile
Price feedsDay-ahead prices pulled in from a feed.
Using Gridcog has been 10x faster than in-house tools. We can iterate quickly, test multiple scenarios, and provide reliable financial results for investors.
Marc Kleindienst · Energy Solutions Engineer, Prosolia EnergyA Gridcog customer, quoted on gridcog.com

Where Gridcog is now

Gridcog reports more than 18,000 projects modelled across 40-plus markets, with customers including Shell, TotalEnergies, Octopus Energy Generation, AGL and PwC. In July 2026 it raised a US$10 million Series A led by ABB.

Those results belong to Gridcog's whole team. My part is the interface their customers work in every day.

18,000+

Projects modelled on the platform, as reported by Gridcog.

40+ markets

Energy markets covered, as reported by Gridcog.

US$10m

Series A in July 2026, led by ABB.

Designing software for experts?

If your users know the domain better than anyone and still fight the tool, the structure is usually what needs redesigning. That is the work I do.

An engineer in a hard hat and hi-vis vest looking out over a solar farm and battery storage containers
Petar Ceklic
Replies within a day

Brief me.

Five quick questions, about 30 seconds.

press Enter ↵

Get in touch

👋 Hello, I live in sunny Perth, Western Australia.

If you've got a project in mind, send me a message and I'll reply within a day. If it makes sense to talk after that, we can book a short chat or grab a coffee in Perth.

Petar Ceklic
Petar Ceklic
Replies within a day

Brief me.

Five quick questions, about 30 seconds.

press Enter ↵
Petar Ceklic
15 minute chat