Agentic AI Use cases

Overnight failures grouped and owned before the stand-up, with freeze in view.

A regression is the batch of tests that runs against the chip design every night, and RTL is the design written as code before it becomes silicon.

WHAT JUNO’S ROLES WOULD DO AROUND THE DESIGN TOOLS01Regressioncompletesnightly about 05:00, or oneach continuous run02Group by signatureknown ones attach to theirbug; a new one on threeseeds opens one03Assign an ownera known signature to itslast owner; a new one tothe block owner04Call the designtool’s agentroot cause and a proposedfix, recorded against thebug05Rerun and reporttests whose bug is fixedrerun; coverage and openbugs on one pageevery reassignment and rejected root cause teaches the next run
The design tools’ agents do the root cause and the fix; Juno’s roles would orchestrate, watch, record and escalate around them. Designed, yet to be built.
60% to 70%
of a chip project’s engineering effort goes to verification (Siemens, 3 September 2025)
dated source
14%
first-silicon success in the 2024 Wilson Research Group study, the lowest in over twenty years (Semiconductor Engineering, 27 March 2025)
dated source
25% to 40%
shorter debug cycle time with an autonomous debug workflow, published by the vendor (Synopsys, 27 July 2026)
dated source
312
failures in nine signatures on one showcase morning, each with an owner before 09:00
showcase estate
A moment in the day

Where this use case starts

VL
Wednesday 07:30 · the 7741-Y interface block · after the nightly regression

The nightly regression on the 7741-Y controller’s interface block finished at 05:10 with 312 failures out of 4,120 tests. By 07:30 Juno has grouped them into nine signatures. A signature is the pattern of the error message and the point in the test where it fails, so two failures with one cause carry one signature. Six signatures match open bugs and are attached to them with the log and waveform references; three are new and have gone to their block owners.

For the largest group Juno called the design tool’s own debug agent, which came back with a root cause in the interface arbiter and a proposed fix; the block owner has both in her queue, with the tool’s evidence beside them. The 240 tests that failed on a seed whose bug was marked fixed yesterday are rerunning.

The verification lead opens one page: coverage at 91.4%, two bugs past their service-level clock, three new signatures waiting for his confirmation. The 09:00 stand-up starts from that page. showcase estate

Why this is a top use case

What the market is doing, with the evidence

The design-tool vendors (EDA, electronic design automation) have moved fastest, and their figures are their own. Synopsys reports an autonomous debug closure workflow that cuts debug cycle time by 25% to 40%, with AMD evaluating it (Synopsys, 27 July 2026) 33. Its orchestrated design and verification workflow doubles productivity, up to five times in some cases (Synopsys, 11 March 2026) 34. Siemens ships a verification planning and debug toolkit used by MediaTek, and an agent endorsed by Samsung and NVIDIA with human checkpoints and audit trails (Siemens, 27 February 2026; Siemens, 16 March 2026) 41, 40; Samsung reports design turnaround down about 50% (Samsung, 17 March 2026) 43.

Cadence reports its ChipStack agent cutting verification effort by about ten times in some areas at Altera and formal verification time by up to four times at Tenstorrent. It produced a test plan in under 40 minutes against a week by hand (EE Times, 10 February 2026) 37. Its virtual engineer with NVIDIA claims a five-week validation loop done in under a day (Business Wire, 31 May 2026) 38.

The pain is measured. Verification takes 60% to 70% of a chip project’s engineering effort (Siemens, 3 September 2025) 32. First-silicon success fell to 14% in the 2024 Wilson Research Group study, the lowest in over twenty years, and 75% of projects ran behind schedule (Semiconductor Engineering, 27 March 2025) 31. McKinsey estimates that AI could nearly double the innovation rate in electronics (McKinsey, 20 June 2025) 49 and, in an earlier semiconductor study, cut R&D cost by 28% to 32% (McKinsey, 2 April 2021) 56.

Where the work sits is the honest part of this page. The design tools’ agents do the root cause and the fixes, inside the simulator and the debugger. Juno’s roles would run the flow around them: start and watch the regression, group and assign the failures, call the tool’s agent and record its answer. They would rerun what a fix should clear, keep coverage and bug convergence on one page and escalate when a clock runs out. The record of who decided what stays on the platform.

By hand today

How it is done today

A nightly regression of thousands of tests runs through scripts and a job scheduler. Verification engineers spend the morning reading logs, grouping failures by hand, many of them duplicates, and assigning them to block owners by e-mail, chat and the bug tracker. Designers debug with waveforms and rerun; a fix that clears one seed is checked on the others by hand.

A weekly coverage review and a bug scrub meeting set priorities, and the programme manager builds status slides for the RTL freeze review from three tools. Ten to fifty engineers and six tools are involved, and the vendors put one verification loop at about five weeks (Business Wire, 31 May 2026) 38.

05:00
The regression finishesThousands of tests through scripts and a scheduler; the logs wait in a directory.
07:30
Reading the logsEngineers open the logs and group the failures by hand; many are duplicates of yesterday’s.
09:00
Assigning by e-mail and chatOwners are guessed from the block, then corrected; the bug tracker is updated later.
All day
Debug and rerunDesigners debug with waveforms and rerun; a fix that clears one seed is checked on the others by hand.
Weekly
Coverage review and bug scrubPriorities are set in a meeting; status slides for the freeze review are built from three tools.
Five weeks
One verification loopThe vendors put the loop at about five weeks (Business Wire, 31 May 2026).

Illustrative times for a morning done by hand; the five-week loop is the vendor’s own figure 38.

Stage by stage

How Juno enables it

Juno would sit beside the design tools, and the six stages below say what it would do at each. The regression manager, the coverage database and the bug tracker are registered and read in place; the design objects join the ontology; the triage policy is written in plain language; the design tools’ own agents are called as tools for the root cause; and every reassignment teaches the next run.

01
Register
The regression manager’s result tables, the coverage database, the bug tracker, the design data management log, the scheduler’s job records and the programme schedule would be registered through the data fabric and read in place. The objects: regression run, test and seed, failure signature, log and waveform reference, coverage point and hole, bug, owner and block, milestone, compute and licence job.
02
Map
The supply chain ontology would be extended with design objects: block, testbench, regression run, test, failure signature, bug, coverage point and milestone, and how they relate. A failure belongs to a run, refers to a test and a seed, and may attach to a bug. Juno proposes the mappings with a confidence and a reason; the CAD team, the engineers who run the design tools and flows, confirms them.
03
Ask
“Which failures are new since last night, and who owns them?” and “Which coverage holes on the interface block are still open past their closure date, and which waivers cover them?” Each answer shows the tables it read, the query and the time it took, with the asking engineer’s access rights applied.
04
Teach
The team’s own words: how signatures are named, the block aliases the design uses in place of their formal names, and the waiver vocabulary. A waiver is an agreed exception for a coverage hole the team will accept at freeze. “Sig”, “seed” and the team’s bug severities would be taught as the team writes them.
05
Act
Trigger: a regression completes, a milestone date arrives, a bug passes its clock or the freeze gate approaches. The policy: “Group failures by signature; assign a known signature to its last owner; open a bug for a new signature seen on three seeds; rerun tests whose bug is marked fixed.” It would be released at A3, with exceptions to the verification lead. Write-back: bugs to the tracker, which returns the issue key; reruns to the scheduler, which returns the job identifier.
06
Learn
Every reassignment, rejected root cause, merged grouping and waiver would be recorded with the rule version and the facts at the time. The measure that improves is failures triaged before the stand-up, with the duplicate-failure share beside it. The assignment rules improve with each release, and drift detection flags a block whose signatures stop matching the ones the rule learned.
1
05:10 · The regression completesThe scheduler reports the run on the 7741-Y interface block finished: 4,120 tests, 312 failures. The trigger fires and the rules engine reads the triage policy, version 2.
312 failures
2
05:12 · Grouped by signatureThe failures fall into nine signatures. Six match open bugs and attach to them with their log and waveform references.
9 signatures
3
05:20 · Three new signaturesEach was seen on three seeds or more, so a bug is opened for each and assigned to the block owner; the tracker returns the issue keys.
3 bugs opened
4
05:25 · The design tool’s agent is calledFor the largest group, 140 failures on one signature, Juno calls the design tool’s debug agent as a tool and waits for its answer.
waiting
5
06:10 · Root cause returnedThe tool proposes a cause in the interface arbiter and a fix. Juno records both against the bug with the tool’s confidence of 0.84 and routes it to the block owner.
0.84
6
06:15 · Reruns submittedThe 240 tests that failed on a seed whose bug is marked fixed are resubmitted; the scheduler returns the job identifiers.
240 reruns
7
07:30 · One pageCoverage 91.4%, two bugs past their service-level clock, three new signatures awaiting confirmation. The verification lead reads it before the stand-up.
read
8
08:40 · The exceptions decidedThe verification lead confirms two of the three new groupings and merges one with an existing bug; the block owner accepts the root cause and edits the fix.
decided
9
09:00 · Stand-upThe team starts from the page and the record: every failure has an owner, and the two overdue bugs are named.
on record

One Wednesday morning as the run trace would record it: the design tool’s agent is called at step 4 and answers at step 5; people decide at steps 7 and 8. Showcase estate; the roles are designed, yet to be built.

Who does what

The person and the agent

The agent A3 Delegated
  • Groups the overnight failures by signature and attaches known ones to their bugs
  • Assigns a known signature to its last owner and opens a bug for a new signature seen on three seeds
  • Calls the design tool’s agent for the root cause and records its answer against the bug
  • Reruns tests whose bug is marked fixed, updates the coverage tracker and publishes the freeze report
the exception queue: new signatures, proposed root causes and clocks that ran out
The person
  • The verification lead confirms groupings and owners for new signatures each morning
  • The block owner accepts or rejects the root cause and fix the design tool proposed
  • The verification lead approves coverage waivers at the weekly review; the CAD manager approves compute above the daily budget
  • The verification lead and the programme manager decide freeze readiness at the gate; an auditor samples the record

Inside the design tools the market’s level runs from execute within rules to autonomous: the three design-tool vendors describe fully autonomous debug closure and orchestrated workflows with engineer oversight, and say they are moving toward full autonomy (2026) 33, 38, 40. The blueprint designs Juno’s write-back roles here at A3: the agent executes triage, assignment and reruns inside versioned rules, and the verification lead handles exceptions only. Regression run orchestration and compute and licence scheduling are designed at A4, RTL freeze readiness at A2 with a named approver, and the freeze decision itself stays with people.

What you would configure

The settings on this agent

Agent settings · Verification regression triage and debug closure
ElementSettingSet byChanges
ObjectiveEvery overnight failure grouped, owned and, where a fix exists, rerun before the 09:00 stand-up; coverage and open bugs in view toward RTL freezeThe verification lead, in a sentenceNew version
Data scopeRegression results, coverage database, bug tracker, design data log and scheduler records for the 7741-Y programmeA. Novak, data lead, with the CAD teamNew version
RulesThe triage policy and its planning table: a known signature to its last owner; a new signature on three seeds opens a bug; rerun on fixedThe verification lead with the Planner AgentNew version, simulated first
TriggerOn completion of the nightly regression, about 05:00, and of each continuous run; milestone dates; a bug’s clock; the freeze gateThe verification leadAt once
Actions permittedGroup and assign · open or update a bug · call the design tool’s agent · submit reruns · update the coverage tracker · publish the freeze reportThe verification lead, ownerNew version
LimitsReruns per night 2,000 tests · compute above the daily farm budget waits for the CAD manager · one rerun per fix · grouping confidence floor 0.80The verification lead and the CAD managerNew version
EscalationA bug past its clock, to the verification lead with its history · a new signature on a freeze-critical block, to the block owner at once · a failed write to the tracker, to the ownerThe verification leadAt once
LevelA3 Delegated with the verification lead on exceptions; freeze readiness at A2 with the programme manager as approverThe engineering directorUp on evidence, down in one step
FeedbackEvery reassignment, rejected root cause and waiver is recorded and used in the next cycleAlways onFixed

Every row is a setting or a versioned rule; the level and the limits are the ones a manager signs.

Built from the blueprint

The roles this use case would draw on

Lima’s agent blueprint is a design of 960 agent roles across five end-to-end chains. It is the map of what Juno can be configured to do, and the first agents are built from it one at a time, the way the PO cancellation agent was.

Blueprint · designed, yet to be builtThese roles are designed on the map. Each is configured from the same parts.
Design 8 roles
Regression run orchestration A4
Verification failure triage A3
Coverage gap analysis & closure A3
Bug convergence analytics A3
RTL freeze readiness A2
Compute & licence scheduling A4
Regression farm capacity planning A3
Design review gate evidence A3

Additional capability in Idea-to-Market lists the rest of the map for this chain.

How it is measured

The measures that show it working

Measure
Debug cycle time, vendor-reported reduction of 25% to 40% (Synopsys, 2026)
Verification loop length, vendor-reported five weeks to under a day (Cadence, 2026)
Coverage, vendor-reported gain of 20% (Synopsys, 2026)
Failures triaged before the morning stand-up
Duplicate-failure share
Bugs open past their service-level clock
Time to RTL freeze
First-silicon success, baseline 14% (Wilson Research Group 2024 study)

Baseline failures triaged before the morning stand-up first. It is counted from the tracker’s timestamps on day one, before any role is built, and it is the number the desk feels: how many of last night’s failures had an owner at 09:00. Debug cycle time and the verification loop length are the vendors’ headline figures and the table marks them as such. Time to RTL freeze and first-silicon success are read once per tapeout, the point where the finished design goes to the foundry.

Terms on this page
RegressionThe batch of tests run against the design every night or on each change; its failures are the morning’s work.
Failure signatureThe pattern of an error message and point of failure that groups many failures as one cause.
SeedThe random starting value a test runs from; the same seed reproduces the same run.
Coverage hole and waiverA part of the design the tests have yet to exercise, and an agreed exception for one.
RTL freezeThe gate after which the design code changes only through an approved change order.
Questions
What do the design tools’ agents do, and what would Juno do?

The tools’ agents find the root cause and propose the fix inside the simulator and the debugger. Juno’s roles would run the regression, group and assign the failures, call those agents as tools, rerun what a fix should clear and keep the record.

Which failures would still reach a person?

Every new signature, every proposed root cause and every bug past its clock. Known signatures are assigned within the rule, and a rerun runs once per fix; the verification lead sees the exceptions in one queue each morning.

Where would the design data sit?

Where it is. The regression results, the coverage database and the bug tracker are read in place through the data fabric, with each engineer’s access rights applied, and only the triage records, bugs and rerun jobs are written back.

How would the level be raised or lowered?

On evidence. The blueprint designs the triage roles at A3 and freeze readiness at A2. A level goes up when recorded cases pass the test and the owner signs; it comes down in one step, at once.

Next in this chain
The question this leaves

The regression found the bug and the block owner fixed it in the RTL. A fix that reaches silicon or a board becomes a change order. What does one change touch across the company?

Answered next

Engineering change impact across BOM, stock and orders

Every assembly, plant, stock position and open order a change touches, found across the product lifecycle system and two ERPs, drafted for the change board.

Sources

Dated sources used on this page

  1. [31] Semiconductor Engineering (reporting the 2024 Siemens EDA and Wilson Research Group Functional Verification Study), First-Time Silicon Success Plummets, 27 March 2025 (checked 23 September 2026) · https://semiengineering.com/first-time-silicon-success-plummets/
  2. [32] Siemens Digital Industries Software (Verification Horizons blog), Why First-Silicon Success Is Getting Harder for System Companies, 3 September 2025 (checked 23 September 2026) · https://blogs.sw.siemens.com/verificationhorizons/2025/09/03/why-first-silicon-success-is-getting-harder-for-system-companies/
  3. [33] Synopsys, Synopsys Advances Agentic AI Chip Design with AMD and Microsoft, 27 July 2026 (checked 23 September 2026) · https://news.synopsys.com/2026-07-27-Synopsys-Advances-Agentic-AI-Chip-Design-with-AMD-and-Microsoft
  4. [34] PR Newswire (Synopsys), Synopsys Outlines Vision for Engineering the Future (Synopsys Converge), 11 March 2026 (checked 23 September 2026) · https://www.prnewswire.com/news-releases/synopsys-outlines-vision-for-engineering-the-future-302711205.html
  5. [37] EE Times, Cadence Unveils ChipStack AI Agent for Agentic Chip Design and Verification, 10 February 2026 (checked 23 September 2026) · https://www.eetimes.com/cadence-unveils-chipstack-ai-agent-for-agentic-chip-design-and-verification/
  6. [38] Business Wire (Cadence), Cadence Unveils Industry's First Fully Autonomous Virtual Engineer for Chip Design, powered by NVIDIA, 31 May 2026 (checked 23 September 2026) · https://www.businesswire.com/news/home/20260531072918/en/Cadence-Unveils-Industrys-First-Fully-Autonomous-Virtual-Engineer-for-Chip-Design-powered-by-NVIDIA
  7. [40] Siemens, Siemens launches Fuse EDA AI Agent, 16 March 2026 (checked 23 September 2026) · https://news.siemens.com/en-us/siemens-fuse-eda-ai-agent/
  8. [41] Siemens, Accelerating IC design and verification with agentic AI (Questa One Agentic AI Toolkit), 27 February 2026 (checked 23 September 2026) · https://news.siemens.com/en-us/questa-one-agentic-ai-toolkit/
  9. [43] Samsung Semiconductor, Samsung Showcases Agentic AI-Driven Semiconductor Engineering Innovation at NVIDIA GTC 2026, 17 March 2026 (checked 23 September 2026) · https://semiconductor.samsung.com/news-events/tech-blog/samsung-showcases-agentic-ai-driven-semiconductor-engineering-innovation-at-nvidia-gtc-2026/
  10. [49] McKinsey & Company (QuantumBlack), The next innovation revolution, powered by AI, 20 June 2025 (checked 23 September 2026) · https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-next-innovation-revolution-powered-by-ai
  11. [56] McKinsey & Company, Scaling AI in the sector that enables it: Lessons for semiconductor-device makers, 2 April 2021 (checked 23 September 2026) · https://www.mckinsey.com/industries/semiconductors/our-insights/scaling-ai-in-the-sector-that-enables-it-lessons-for-semiconductor-device-makers

See Juno answer and act on your own data

Bring two systems and one decision your planners make every week. In the first week we register the data, confirm the mapping with your team, answer your questions in ChatLima and run the first agent at the level you choose.