Case study 01 GSA / IT Vendor Management Office
From Findable to Answerable
Rebuilding the information architecture for federal IT buyers, then designing the assistant that picked up where navigation ran out.
- Client
- GSA / IT Vendor Management Office
- Role
- Research, information architecture, design standards, conversational AI
- Timeline
- Redesign three months, evaluation through November 2023
- Team
- One designer, one developer, program stakeholders
Context
Everything was published. Nothing was findable.
The IT Vendor Management Office helps federal buyers make sense of IT acquisition. Contracting officers, program managers, and small-business specialists arrive with real decisions in front of them: which vehicle covers a given service, whether a set-aside applies, what a pricing escalation clause permits. This ran in two parts. An information architecture rebuild, which is where the friction numbers below come from, and a source-constrained assistant for the questions navigation could not answer, which reached beta and stopped there.
The information was all there. 221 resources across seven topic areas. Technically, everything a buyer needed was on the site.
Content inventory
221 resources, seven topic areas
Counts exceed the total because resources carry more than one tag.
The problem
A published document cannot answer a buyer’s question.
The site was organized around how the agency filed its content, not around what a buyer was deciding. Finding an answer meant knowing which of seven categories it lived under, and buyers do not think in those categories.
In federal acquisition, a confidently wrong answer is worse than no answer.
A buyer who acts on invented guidance about procurement rules has a compliance problem, not a usability problem. That ruled out the obvious approach before it started.
Constraints
What shaped every decision that follows.
- Three months to redesign, evaluation ran longer
- Against an existing live site.
- One developer
- Every decision had to be buildable by a single person inside the timeline.
- Non-technical maintainers
- Business analysts and project managers needed to update content afterwards without a developer, so the CMS was a design constraint rather than an implementation detail.
- Section 508 and WCAG
- Applied to a conversational interface where the patterns are less settled than they are for pages.
- No invented answers
- The system could not reach past what it had been given.
Discovery and research
Two inputs, pulling in the same direction.
Google Analytics gave the behavioral picture: where buyers entered, what they searched for, where sessions ended without a result. It showed what was happening but not why. So I built the qualitative side alongside it: personas, journey maps, and a full architecture audit of what existed against what buyers were trying to do.
Buyer journey
The site got them close. Close was not the answer.
-
Step 1
Arrives with a specific question
A vehicle, a clause, a set-aside. One decision, already in mind.
-
Step 2
Task-based navigation gets close
The redesigned architecture lands them on the right topic in a click or two.
-
Step 3, stall point
The resource is close, not exact
Adjacent guidance. A different vehicle, year, or clause.
-
Step 4
Re-reads, hedges, or gives up
No way to confirm “close enough” without finding a person to ask.
Read together, the two agreed. Analytics showed sessions ending on category pages. The journeys showed why: a buyer arrives with a decision, and the site asked them to already know which of seven filing categories the answer lived under.
To measure it I used the PURE method, a structured expert evaluation that scores each task by the difficulty a qualified user would face. It suits situations where direct user access is limited, which this was: federal buyers are hard to recruit and three months did not allow for a full study. I scored seven representative buyer tasks three times. A baseline in June, a mid-build check in August, and a final round against the redesigned site in November.
PURE gave me expert judgment about how hard a task should feel. Analytics showed what people did. The two disagreed in useful places: content that scored acceptably still showed heavy drop-off, which usually meant buyers were arriving from search and landing mid-hierarchy with no orientation.
PURE evaluation
Seven buyer tasks, three rounds of friction scores
Seven-task PURE scorecard comparing June 29, August 15, and November 7, 2023. Round totals are 77 in June, 53 in August, and 25 in November. Every round shows the per-task total and one colored segment for each scored step.
| Buyer task | June June 29, 2023 77 | August August 15, 2023 53 | November November 7, 2023 25 |
|---|---|---|---|
| Request a contract review via the Contract Review Service |
June: 10 points across 5 steps.
|
August: 6 points across 4 steps.
|
November: 5 points across 3 steps.
|
| Use the GW Procurement Equity Data Tool |
June: 11 points across 6 steps.
|
August: 9 points across 6 steps.
|
November: 3 points across 2 steps.
|
| Access an OEM webinar |
June: 5 points across 3 steps.
|
August: 3 points across 3 steps.
|
November: 2 points across 2 steps.
|
| Find technical standards for a new acquisition |
June: 14 points across 6 steps.
|
August: 11 points across 7 steps.
|
November: 3 points across 2 steps.
|
| Discover small business opportunities |
June: 18 points across 7 steps.
|
August: 10 points across 6 steps.
|
November: 8 points across 5 steps.
|
| Read about the 2024 ITVMO Summit |
June: 4 points across 4 steps.
|
August: 3 points across 3 steps.
|
November: 1 points across 1 steps.
|
| Learn about the newly BIC-designated Polaris contract |
June: 15 points across 6 steps.
|
August: 11 points across 7 steps.
|
November: 3 points across 3 steps.
|
Each column is a round, each row a task, and each segment one scored step. Green is low friction, amber moderate, red high. The task totals add up to 77 in June, 53 in August, and 25 in November; the step count falls from 37 to 36 to 18 because the redesign removed steps, not only friction.
Task Friction Score Reduction
Comparison of June baseline and November final PURE scores for seven buyer tasks. ITVMO Summit resources decrease from 4 to 1, OEM webinar from 5 to 2, procurement equity data tool from 11 to 3, finding technical standards from 14 to 3, Polaris contract from 15 to 3, contract review or RFP from 10 to 5, and discovering small-business opportunities from 18 to 8.
-
ITVMO Summit resources
-75% -
OEM webinar
-60% -
Procurement equity data tool
-73% -
Finding technical standards
-79% -
Polaris contract (BIC-designated)
-80% -
Contract review or RFP
-50% -
Discovering small-business opportunities
-56%
PURE friction score per task; lower is better. The comparison isolates the June baseline and November final scores; the August checkpoint remains in the matrix above.
How the work was made
Foundations, then fidelity, then sign-off.
-
Personas and journeys
I mapped how contracting officers, program managers, and small-business specialists move from a question to a decision. Those journeys became the test for every structural decision: if a proposed architecture did not shorten one of them, it did not go in.
Buyer journey
Choose a compliant IT vehicle
Journey map for choosing a compliant IT vehicle across five stages: define need, find options, compare fit, verify authority, and decide or recover. Confidence scores out of five are three, two, three, four, and five.
-
Stage 01
Define need
“What am I buying?”
Confidence score: 3 out of 5. -
Stage 02
Find options
“Which vehicles could work?”
Confidence score: 2 out of 5. -
Stage 03
Compare fit
“What tradeoffs matter?”
Confidence score: 3 out of 5. -
Stage 04
Verify authority
“Can I defend this guidance?”
Confidence score: 4 out of 5. -
Stage 05
Decide or recover
“Do I have enough evidence?”
Confidence score: 5 out of 5.
Confidence dips while comparing look-alike options, then climbs once a claim can be checked against a cited source.
One of the journeys mapped during discovery, redrawn here. Confidence dips while comparing look-alike vehicles, then climbs once authority can be checked against a cited source. -
-
Site architecture
The architecture came out of the journeys rather than the existing content taxonomy. That is the whole move in this project. The old structure described how ITVMO files things; the new one describes what a buyer is trying to decide.
-
Low fidelity first
I stayed in wireframes for as long as the questions were structural. A stakeholder looking at a gray wireframe argues about whether the structure is right; a stakeholder looking at a polished screen argues about the shade of blue.
-
Prototype as specification
Once structure settled I moved to high-fidelity interactive prototypes in Axure, so the developer could see behavior rather than infer it from a static comp. On a compressed timeline the round trip is the expensive part.
-
Stakeholder review as the gate
Nothing was built before program stakeholders approved it. With one developer and three months, a late objection would have been unrecoverable, so clearing each stage before moving on was what made the timeline hold.
What I tried first
The simpler fix, because it was clearly broken.
I re-architected navigation and wayfinding around buyer tasks instead of agency filing categories, built a content-pattern library so every resource type had a consistent shape, and wrote plain-language guidance for the people producing content.
Information architecture
Filed by category, or organized by task
Before
The buyer must know which of seven categories the answer lives under.
- Home
- Resources
- Topic area
- Filter set
- Resource
Five steps, and step three requires a guess.
After
The buyer states the decision they are making and is routed to the answer.
- Home
- Task path (key step)
- Answer
Three steps, and no category knowledge required.
From filing categories to buyer tasks
Information architecture comparison. Before: seven equal sibling categories under Home, including acquisition best practices, market intelligence, contract solutions, policy, technology, small business, and ITVMO general. After: three task-based groups under Home—Decide and buy, Learn and compare, and Ask and participate—with three destinations inside each group.
Before — filed by category
- Acquisition best practices
- Market intelligence
- Contract solutions
- Policy
- Technology
- Small business
- ITVMO general
Seven equal siblings. A buyer has to know which one holds the answer before they can start.
After — organized by task
Decide and buy
- IT vehicles
- Services
- Small-business support
Learn and compare
- Resources
- Community
- Related guidance
Ask and participate
- Ask Vemo
- Human escalation
- Contact
Three destinations named for the decision a buyer is making, not the office that filed the content.
We built on Netlify with a git-based CMS so business analysts and project managers could publish events and add resources themselves, with no ticket and no developer in the queue. Without a content-pattern library, anyone who could publish anything would recreate the original problem inside a year. The patterns constrained publishing into shapes the architecture already accounted for, and I ran the training that got the team using them.
Standards
Fixing the site was not enough on its own.
ITVMO publishes fact sheets, slide decks, and reports as well as web pages, and all of it was drifting. A buyer who found a consistent answer on the site and then downloaded an inconsistent PDF was back where they started.
So I authored the ITVMO Document Style Guidelines: a logo system with spacing rules and misuse examples, 508 color-contrast guidance with pass and fail cases, a three-typeface hierarchy, primary and secondary palettes with hex values and shade ramps, and full specifications for fact sheets and slide decks down to margins, line spacing, and paper weight.
The constraint
It solved one problem and created another.
Even a well-organized hierarchy still made buyers translate their question into someone else’s structure. That pointed toward a conversational layer, and in federal acquisition a chatbot is a liability as much as a feature.
The system was restricted to retrieve only from an approved document set. It could not reach past what it had been given, and could not generate an answer that was not grounded in a source. That solved accuracy completely and created the design problem that defined the work.
A source-constrained system says it does not know far more often than a general-purpose one. How it handles that moment determines whether anyone uses it twice.
Four states, and only one is the clean answer.
-
The grounded answer
Every response carries its source, visible before the user acts. A contracting officer putting their name on a decision needs to check the guidance, and a citation they have to hunt for is a citation they will not use.
-
The partial answer
Sources cover part of the question. The system says which part, rather than filling the gap or refusing entirely.
-
The gap
No approved source means no guess. The system says so and routes to a person, rather than filling the gap.
-
The recovery
A wrong or partial answer keeps what the user already established. Recovering does not mean starting over.
Underneath all four: latency. Retrieval takes time, silence reads as broken, and a spinner reads as stalled. The system had to show what it was doing.
An empty field is a hard interface.
The assistant opens with four routes: find events, explore resources, learn about ITVMO, watch past events. Alongside them sit suggestions drawn from buyers’ recurring questions. That reduced the blank-input problem and set expectations about scope before anyone hit a wall.
The framework
The trust patterns generalized past this project.
I co-authored a responsible-use framework with ethics, compliance, and risk, covering transparency, latency feedback, error handling, and user-trust design across conversational interfaces, automated guidance, and AI-assisted navigation. Built so the next team would not relitigate the same questions.
Responsible-use framework
Four principles, one standard everywhere.
Co-authored with ethics, compliance, and risk. Applied identically across all three surfaces.
-
Transparency
Where an answer came from, visible before a user acts on it.
-
Latency feedback
What the system shows while it is working, so silence does not read as broken.
-
Error handling
What happens when a source does not cover the question: no invented answers, no dead ends.
-
User-trust design
How recovery works after a wrong or partial answer, so one miss does not end the session.
- Conversational interfaces
- Automated guidance
- AI-assisted navigation
What came of it
Friction fell 68%, and the assistant never shipped.
Measured outcome
Task friction fell 68% across three evaluations
PURE method. Lower is better. Each round scored the same seven buyer tasks.
- 77 PURE friction score June 29, 2023
- 53 PURE friction score August 15, 2023
- 25 PURE friction score November 7, 2023
Every one of the seven tasks improved. The largest movers were finding technical standards for a new acquisition, which fell from 14 to 3, and learning about the newly BIC-designated Polaris contract, which fell from 15 to 3. The step counts fell too. Across the seven tasks, 37 steps in June became 36 in August and 18 in November. August cut friction inside the existing steps; November removed the steps.
Shipped
- Information architecture organized around buyer tasks
- Content-pattern library and plain-language guidance
- Git-based CMS the program team publishes to without a developer
- ITVMO Document Style Guidelines, twenty-four pages
Did not ship
- The assistant was completed to beta and never deployed
- The budget for it did not survive, a funding decision rather than a design one
The retrieval-constrained approach matched what the industry now calls RAG. I was designing for it before the vocabulary settled. The architecture matched the pattern, not that I built a RAG system to current standards.
What I would do differently
Test uncertainty, separate the scope, finish the hardest task.
-
I would have tested the uncertainty states specifically
I designed them from principle and from what the retrieval constraint implied, and I believe they were right. But I never watched a contracting officer hit “I don’t have an approved source” and decide whether to trust the system afterward. That is the moment the whole design turns on, and I have reasoning for it rather than evidence.
-
I would have separated the two projects earlier
The IA work and the conversational layer were treated as one effort because they shared a goal, and the assistant’s funding got tied to a scope it did not need. Split, the assistant might have found a smaller budget line of its own.
-
One task never got where it should have
Discovering small business opportunities started worst at 18 and ended highest at 8. Every other task finished at five or below. I ran out of runway before I solved that one, and it is the task with the most public-policy weight attached to it.