STED
2026
81% of project professionals say projects have become more complex, while only 41% of agency teams trust that things don’t slip through the cracks.
STED is a pre-project decision workspace for small agencies that brings delivery expertise into the decision earlier — surfacing uncertainty before it becomes a client commitment.
Problem
When rough estimates become client promises
Small agencies often need to estimate work before designers, engineers or other specialists have reviewed what it involves.
An unchecked integration, unclear requirement or assumed availability can turn a rough estimate into an unrealistic commitment, putting timelines, resources and margins at risk.
I initially explored how agencies could estimate faster. But desk research and a review of existing tools shifted my focus: the bigger concern was unresolved questions being treated as facts.
Sales tools help teams win work. Delivery tools help them manage it. In the tools I reviewed, the decisions between those stages received less attention.
What mattered wasn’t simply whether information was missing. It was whether the team knew what was still uncertain before acting on it.A timeline might depend on an integration being possible. An estimate might assume specialist capacity is available. A requirement might still need technical validation. An unresolved question becomes dangerous when it starts behaving like a fact.
Rationale
Why the Delivery Lead?
The Account Lead understands the client, specialists assess what the work requires, and the founder considers the commercial implications.
I chose the Delivery Lead as the primary user because they bring those perspectives together to assess what the team can realistically deliver.
The design question became:
How might we help agencies make uncertainty visible before committing to a project?
The Delivery Lead has to bring those perspectives together and assess what the team can realistically deliver.
Decisions
Three decisions shaped STED
01 — Make uncertainty explicit
Not every unknown needs an answer before work can move forward.
Rather than simply flagging missing information, I designed STED to distinguish between questions that need resolving, assumptions the team is making and risks it chooses to accept.
The potential impact on time, cost or resources remains visible, helping teams understand what they're carrying into a commitment.
02 — Design for changing information
A fixed eight-step journey didn't reflect how opportunities evolve.
The initial interface direction followed a linear sequence. But new client information or specialist input can change an earlier decision at any point.
I replaced that sequence with persistent navigation, allowing teams to revisit an opportunity as new information emerges without restarting the process.
03 — Make AI findings reviewable
AI should highlight potential risks, not make decisions on the team's behalf.
Generic findings were difficult to assess, so I designed each flag to explain why it was raised and show its source where possible.
I also changed the system language from “Requires clarification” to “Needs your review”, supported by Accept, Edit and Remove actions.
This makes human review an explicit part of the proposed workflow rather than treating AI output as an established fact.
Process
Using AI without handing over the design decisions
I defined STED's roles, decision ownership and AI boundaries before exploring the interface. I also established its visual foundations in Figma using variables and reusable patterns.
I then used Claude Design to generate an initial direction and evaluated it against the product model.
That review led me to reject the fixed journey, generic AI findings and overly authoritative language. I refined the resulting interface in Figma.
The value of AI was in accelerating exploration; the product decisions came from evaluating what did and didn't fit the problem.
Reflection
What STED still needs to prove
STED is a product hypothesis, not a shipped product.
My first priority would be to test whether the proposed decision-making workflow reflects how agencies actually operate, particularly how Delivery Leads, Account Leads and founders share responsibility.
I'd then explore whether STED helps teams make clearer commitments without slowing down the process.
The measures I'd investigate include estimate accuracy, unexpected resourcing, scope changes and surprises after kickoff.
Longer term, comparing early assumptions with actual delivery outcomes could help agencies learn from previous projects and make better-informed decisions next time.