See what your R&D spend actually bought.
Your R&D budget goes to people and to AI. R&D Insights shows where both went by project, next to the priorities you set for the year.
From the record to the return on every project.
Where the R&D budget went, project by project.
Every engineering hour and AI token lands on the project it built. Nobody fills in a time sheet.
AllocationAI spendPrioritiesReturn
R&D spend by project106 people · $20.2M
ProjectShareSpend
- $7.9MPayments ledger servicePlatform · 38 people
- $5.9MFraud model v3Risk ML · 21 people
- $3.0MMobile onboardingGrowth · 26 people
- $2.6MSupport agentCX · 12 people
- $0.8MAdmin toolingIT · 9 people
From Jira, GitHub, payroll and your AI providers. Nobody fills in a time sheet.
AI spend next to engineering time.
Token spend lands as one invoice. R&D Insights breaks it out by the projects that used it. A project's cost then includes both.
AllocationAI spendPrioritiesReturn
AI usage · October31% of R&D spend is AI
Anthropic API$158,300- OpenAI API$216,700
- AWS Bedrock$116,700
- Cursor$4,240
Cost by project, FY2025EngineeringAI
- Payments ledger service24% AI
- Fraud model v344% AI
- Mobile onboarding10% AI
- Support agent54% AI
- Admin tooling13% AI
Every API key and seat on a project. Seat tools split by commits.
Your stated priorities next to your actual ones.
Neo.Tax shows how much of each person's time went to the priorities you set. It shows how much went elsewhere too.
AllocationAI spendPrioritiesReturn
Stated vs actualPlanActual
PlanProjectActual
- 1Fraud model v3$5.9M spent2nd
- 2Mobile onboarding$3.0M spent3rd
- 3Support agent$2.6M spent4th
- 4Payments ledger serviceOutspending 3 higher priorities1st
- 5Admin tooling$0.8M spent5th
The plan you set, next to where the time and tokens went.
Cost and return on every project.
Spend per project sits next to the revenue it drove. The credit and the capitalized cost sit on the same row.
AllocationAI spendPrioritiesReturn
Cost and return by projectFY2025
ProjectCostRevenueReturn
- Payments ledger service$7.9M$27M3.4×
- Fraud model v3$5.9M$12.4M2.1×
- Mobile onboarding$3.0M$13.8M4.6×
- Support agent$2.6M$4.7M1.8×
- Admin tooling$0.8Mn/an/a
3.4×Payments ledger service cost $7.9M and drove $27M of revenue. It also carries $6.4M of QREs and $1.8M capitalized.
Revenue by project comes from your plan. Everything else comes from the record.
Built on the record tax and accounting already trust. Finance doesn't start a second system.
ProjectsFY2025 · 109,100 hours · 106 people
- Engineering time
- $13.9M
- AI spend
- $6.3M
- Qualified research
- $14.9M
- Capitalized
- $3.9M
- Return on R&D
- 3.0×
Project
Engineering time
AI spend
Tax
§41 · Form 6765
Accounting
ASC 350-40
FP&A
Return · plan
- Payments ledger servicePlatform · 38 peopleEngineering time46,200 h$6.0MAI spend$1.9MTax$6.4MQualifiedAccounting$1.8MCapitalizedFP&A3.4×Plan 4 · Spend 1Over plan
- Fraud model v3Risk ML · 21 peopleEngineering time25,600 h$3.3MAI spend$2.6MTax$4.9MQualifiedAccounting$0ExpensedFP&A2.1×Plan 1 · Spend 2
- Mobile onboardingGrowth · 26 peopleEngineering time21,400 h$2.7MAI spend$0.3MTax$1.9MQualifiedAccounting$1.2MCapitalizedFP&A4.6×Plan 2 · Spend 3
- Support agentCX · 12 peopleEngineering time9,800 h$1.2MAI spend$1.4MTax$1.7MQualifiedAccounting$0.9MCapitalizedFP&A1.8×Plan 3 · Spend 4
- Admin toolingIT · 9 peopleEngineering time6,100 h$0.7MAI spend$0.1MTax$0Not qualifiedAccounting$0ExpensedFP&AInternalPlan 5 · Spend 5
Questions finance asks first.
Is this engineering observability?
Neo.Tax leaves bug counts and cycle time to engineering tools. R&D Insights answers
finance's question: where did the money go.
Do engineers have to track time?
No. It reads the same signals as the R&D credit.
Is it a separate product?
It runs on the same record. If you already use Neo.Tax for tax or accounting it reads
data you've already connected.