A
Agentic AIAI
Agentic AI describes AI systems that plan, take multi-step actions, and call tools in pursuit of a stated goal, rather than producing a single response and stopping. An agentic system can decide what to look up, which tool to invoke, how to verify its own output, and when to ask for help. Production agentic systems usually run on top of a large language model, with an orchestration layer that handles memory, tool routing, and guardrails. Most enterprise use cases today combine agentic AI with strong human review at decision points.
How Agile Visa uses this term: in the AI-Native Agile Operating Model, agentic AI is the engine behind the Augmented Sprint, where agents draft work and humans direct, verify, and decide.
Agentic Sprint ReviewAgile Visa term
An Agentic Sprint Review is a sprint review where AI agents present demos, summarise telemetry, and flag risk signals before the humans take over the conversation. The format keeps the human-led review intact, but shifts the first ten minutes of routine status reporting onto an agent that already has full context. Stakeholders see a tighter narrative, the team spends less time on slides, and the Product Owner can focus the discussion on what changed and what to do next sprint.
How Agile Visa uses this term: we coach teams to adopt the Agentic Sprint Review as the first low-risk step into agentic ceremonies, since the human accountability does not change.
See also: Agentic AI, Augmented Sprint
AI Agile CoachAgile Visa term
An AI Agile Coach is a coach who blends classical Agile coaching stances with AI literacy, model evaluation, and agent-orchestration practice. The role goes beyond running ceremonies. The coach helps a team pick the right model for a job, design evals, set up safe agent loops, and rebuild the working agreements so humans and agents collaborate well. Strong AI Agile Coaches stay current with tooling, but they coach the system, not the tool stack.
How Agile Visa uses this term: AI Agile Coach is the practitioner persona at the core of our AI-Native Agile programs and our founder-designed cohorts.
AI Agile Operating ModelAgile Visa term
The AI Agile Operating Model is a codified system of roles, ceremonies, artifacts, and metrics that allows a delivery team to plan, build, and govern alongside AI agents. It answers the practical questions a leader asks when adopting AI inside delivery. Who decides what an agent is allowed to do. Where do prompts and evals live. How does the team measure leverage. What changes in the Definition of Done. A working operating model gives teams a stable footing while tools and models keep changing underneath them.
How Agile Visa uses this term: this is the umbrella under which our cohort programs, diagnostics, and consulting work are organised.
See also: AI-Native Agile, Model-Aware Backlog
AI CoachRole
An AI Coach helps individuals and teams adopt AI tools deliberately, with attention to skill, governance, and outcomes. The role is broader than prompt training. A good AI Coach maps where AI actually helps in someone's work, sets a learning rhythm, and stops a team from drifting into unsafe or low-value use. AI Coaching is increasingly common inside large enterprises that want adoption without losing control.
How Agile Visa uses this term: distinct from AI Agile Coach. The AI Coach focuses on individual and functional adoption. The AI Agile Coach focuses on delivery teams and the operating model.
AI Coaching StanceAgile Visa term
The AI Coaching Stance is the posture a coach adopts when guiding humans and AI agents at the same time. Classic Agile coaching has stances such as teaching, mentoring, facilitating, and coaching. The AI Coaching Stance adds two more. Verifier, where the coach helps the team check what an agent produced. Capability builder, where the coach uses the moment to lift the team's own skill, not just clear the work. Choosing the right stance per situation is what stops AI use from becoming dependency.
How Agile Visa uses this term: every AI Agile Coach cohort runs a stance lab to practise the full set.
AI EngineerRole
An AI Engineer builds production systems around foundation models, including prompts, tools, retrieval, evaluations, and observability. The role sits between a traditional software engineer and an ML engineer. AI Engineers spend a large share of their time on prompt design, eval harnesses, retrieval pipelines, and guardrails, rather than on training models from scratch. Most enterprise AI roadmaps today need more AI Engineers than ML Engineers.
How Agile Visa uses this term: when we coach delivery teams, we pair AI Engineers with AI-Augmented Product Owners so build and discovery move together.
AI GovernanceAI
AI Governance is the set of policies, roles, and controls that decide how AI is built, bought, deployed, and monitored inside an organisation. Good governance answers concrete questions. Who can approve a new AI use case. What data can be sent to which model. How are model changes reviewed. What is logged. Governance done well is enabling, not blocking. It gives teams a clear path to ship, which is far more productive than vague nervousness about AI.
How Agile Visa uses this term: governance is one of the four pillars in our AI Maturity Diagnostic, alongside people, process, and platform.
AI Maturity DiagnosticAgile Visa term
The AI Maturity Diagnostic is a structured assessment of an organisation's AI delivery readiness across people, process, platform, and policy. It uses interviews, artefact review, and a short survey to produce a maturity score per pillar, plus a sequenced action list. The diagnostic is intentionally light. The goal is not a giant report. The goal is the next three things a leadership team should fund this quarter to move maturity forward.
How Agile Visa uses this term: we run the AI Maturity Diagnostic as the entry engagement for most enterprise clients.
AI-Native AgileAgile Visa term
AI-Native Agile is an operating model that re-architects Agile delivery so that AI agents, prompts, and model outputs are first-class members of the team, not bolt-on tools. It assumes that within a sprint, agents will draft, review, test, and decide on a non-trivial share of the work. The model defines new accountabilities, new artefacts such as the Model-Aware Backlog, and new metrics such as Hybrid Velocity. It keeps human accountability for outcomes intact, while letting teams compound leverage sprint over sprint.
How Agile Visa uses this term: AI-Native Agile is the core idea behind every cohort, diagnostic, and coaching engagement we run.
See also: AI Agile Operating Model, Augmented Sprint
AI Product ManagerRole
An AI Product Manager owns AI-powered features end to end, from problem framing to model selection to live evaluation. The role adds a layer on top of classical product management. The PM must judge model behaviour, design evaluation criteria, decide when to use retrieval versus fine-tuning, and lead the team through fast model and tooling churn. Strong AI PMs do not pretend to be researchers. They build enough literacy to make crisp product decisions.
How Agile Visa uses this term: AI Product Manager is one of the most requested role transitions in our Singapore and Malaysia cohorts.
AI SafetyAI
AI Safety is the field that studies how to keep AI systems aligned with intended behaviour and free from harmful failure modes. It spans concrete engineering work, such as guardrails, evals, and red teaming, and broader questions about long-run alignment. For most enterprises, applied AI Safety today is mostly about three things. Preventing data leakage. Catching unsafe outputs. Logging enough to debug what happened.
How Agile Visa uses this term: we treat AI Safety as a hands-on engineering practice, not a theoretical aside, in our AI Engineer track.
AI-Augmented Product OwnerAgile Visa term
An AI-Augmented Product Owner is a Product Owner who uses AI agents for backlog grooming, prioritisation drafts, persona research, and stakeholder synthesis. The accountability for value does not change. What changes is the speed at which the PO can move from signal to a refined backlog. The same agent loop that drafts a refined story also drafts the eval criteria for that story, which keeps the Definition of Done honest.
How Agile Visa uses this term: this is the target role we coach incumbent POs into during AI-Native Agile programs.
See also: Model-Aware Backlog
Agile CoachRole
An Agile Coach develops the people, teams, and leaders of an organisation so they can deliver value adaptively. The role is wider than Scrum Master. It usually covers multiple teams, leadership coaching, and organisational design. Agile Coaches are most useful where a team or business unit faces a real change of context, since their job is to help the system learn, not to enforce a framework.
How Agile Visa uses this term: many of our learners progress from Scrum Master to Agile Coach to AI Agile Coach.
Allowable Cost MatrixFunding MY
The Allowable Cost Matrix is the HRD Corp reference table that defines which training cost items can be claimed and at what cap. It covers fees, trainer cost, daily allowances, accommodation, and consumables. The matrix is updated periodically, and the applicable version is the one in force at the date of the training. Employers planning HRD Corp claimable programs should map their budget to the matrix before submission, not after.
How Agile Visa uses this term: we share the current Allowable Cost Matrix mapping for every HRD Corp claimable program at quote stage.
Augmented SprintAgile Visa term
An Augmented Sprint is a sprint in which AI agents contribute discovery, drafting, code, test, and review work alongside humans, with humans owning decisions. The sprint cadence stays familiar. Planning, daily, review, retrospective. What changes is the work mix. A meaningful share of the sprint output is drafted by agents, then verified and decided on by humans. The team measures this with Hybrid Velocity to keep capacity and leverage honest.
How Agile Visa uses this term: the Augmented Sprint is the unit of practice in the AI-Native Agile Operating Model.
See also: Hybrid Velocity, AI-Native Agile
B
BacklogAgile
A backlog is an ordered, living list of everything the team might do, refined continuously by the Product Owner. A good backlog is small enough to be useful, deep enough to reflect strategy, and explicit about value and risk. Backlogs go wrong when they become wish lists, when they hide assumptions, or when items at the top are not yet ready for the team to take on.
How Agile Visa uses this term: in AI-Native Agile, we extend the backlog into a Model-Aware Backlog that captures prompts, evals, and model fallbacks.
See also: Model-Aware Backlog, Product Owner
C
Continuous Capability LoopAgile Visa term
The Continuous Capability Loop is a learning rhythm that pairs each sprint with a small, scoped capability uplift so teams keep pace with shifting AI tools. The loop has three beats. Spot a capability gap in the current sprint. Choose one focused practice or asset to build. Carry it into the next sprint and measure whether it stuck. The point is not to be perpetually in training. The point is to make learning a normal sprint output.
How Agile Visa uses this term: every cohort closes with a Continuous Capability Loop plan that the learner runs back at work.
D
Definition of DoneAgile
The Definition of Done is the shared, explicit standard a backlog item must meet to be considered complete and releasable. It is a quality contract inside the team and a transparency tool outside it. Good Definitions of Done evolve. As the product, platform, or regulation changes, the team raises the bar. In AI-Native delivery, the Definition of Done usually grows to include eval thresholds, prompt-asset checks, and fallback verification.
How Agile Visa uses this term: we coach teams to add a one-line AI clause to their Definition of Done before changing anything else.
F
Fine-tuningAI
Fine-tuning is the practice of adapting a pre-trained model to a narrower task or style by continuing training on a smaller, curated dataset. For most enterprise use cases, retrieval-augmented generation and good prompting solve the problem more cheaply than fine-tuning. Fine-tuning earns its cost when you need a very specific style, a very tight output format, or strong behaviour on tasks that prompting cannot reliably steer.
How Agile Visa uses this term: in our AI Engineer track, we teach a decision tree. Try prompting, then retrieval, then fine-tuning, in that order.
See also: RAG, Prompt Engineering
Founder-Designed CohortAgile Visa term
A Founder-Designed Cohort is a small learner group where the founder designs the curriculum, leads the room, and personally signs off on participant artefacts. The format trades scale for depth. Group size is kept small so the founder can call out poor reasoning, push on real workplace context, and personalise the capability plan. The format works best for senior practitioners who already have the basics and need calibrated coaching at the next level.
How Agile Visa uses this term: our AI Agile Coach and AI Product Manager cohorts are Founder-Designed Cohorts led by Prashant Shinde.
H
HRD Corp SBL-KhasFunding MY
HRD Corp SBL-Khas is a Malaysian claimable scheme where HRD Corp pays the training provider directly from the employer's levy account, instead of the employer paying upfront and reclaiming later. The scheme reduces cash-flow friction for employers and removes a common operational reason for skipping good training. The program, trainer, and provider must all meet HRD Corp eligibility for SBL-Khas to apply.
How Agile Visa uses this term: most of our Malaysia enterprise programs are quoted with the HRD Corp SBL-Khas pathway built in.
See also: HRDF Levy, Allowable Cost Matrix
HRD Corp Train-the-TrainerFunding MY
HRD Corp Train-the-Trainer is a Malaysian certification that qualifies an individual to deliver HRD Corp claimable training programs. Holders are recognised by HRD Corp as accredited trainers, which is a precondition for programs to be claimable under most HRD Corp schemes. The certification reflects assessed delivery and design capability, not just attendance.
How Agile Visa uses this term: Prashant Shinde is an HRD Corp Train-the-Trainer certified founder, which is why our Malaysia programs can be claimable end to end.
HRDF LevyFunding MY
The HRDF Levy is a statutory training levy that registered Malaysian employers pay into a fund used to subsidise their workforce training. The levy sits in the employer's HRD Corp account and can be drawn down through claimable schemes. Unused balances are a missed opportunity, since they cannot be redirected to other uses. Most mature finance teams treat the levy as a budget that must be spent on real capability work.
How Agile Visa uses this term: we help client HR and finance teams sequence programs across the year to fully use their HRDF Levy.
See also: HRD Corp SBL-Khas
Hybrid VelocityAgile Visa term
Hybrid Velocity is a throughput measure that separates human story points from AI-completed work units, so a team can see leverage without distorting capacity. Reporting a single inflated velocity number once agents start contributing is misleading. Hybrid Velocity keeps the human side comparable to past sprints, while a parallel agent line shows how much extra value the team is now able to ship. It also makes the cost of an agent's mistakes visible, since rework is tracked as well.
How Agile Visa uses this term: Hybrid Velocity is the metric we coach delivery teams to adopt before any executive dashboard sees agent throughput.
I
IBF-STSFunding SG
IBF-STS is the Institute of Banking and Finance Standards Training Scheme, a funding scheme that supports talent development in the Singapore financial sector. Eligible programs that meet IBF accreditation can be partly funded for individual or company-sponsored learners, with subsidy rates depending on the learner's citizenship and age. Approved trainers and curricula go through IBF accreditation before learners can claim.
How Agile Visa uses this term: selected enterprise programs for SG banks are designed with IBF-STS eligibility in mind.
ICAgile Member OrganizationAgile Visa term
An ICAgile Member Organization is a training provider formally accredited by ICAgile to deliver certified learning programs and award ICAgile certifications. Member Organizations meet ICAgile's design and delivery standards, and their trainers are individually assessed. Learners pay a single, transparent fee that covers both the program and the certification, since the Member Organization is the entity that issues the credential.
How Agile Visa uses this term: Agile Visa is an ICAgile Member Organization, which is why our ICAgile certifications are issued directly by us as the accredited provider.
K
KanbanAgile
Kanban is a flow-based method that visualises work, limits work-in-progress, and improves the system by managing flow rather than fixed iterations. Teams using Kanban map their workflow, set explicit WIP limits per stage, and use measures such as lead time and flow efficiency to spot blockers. Kanban is often a better fit than Scrum for support, platform, and operations teams whose work does not arrive on a sprint cadence.
How Agile Visa uses this term: in our ICAgile catalogue, Kanban-based delivery is taught explicitly as one of several valid Agile flavours.
L
LLM OrchestrationAI
LLM Orchestration is the layer that routes prompts, manages tool calls, handles retries, and stitches together multi-step workflows that run on top of language models. Orchestration is where most production AI complexity lives. The orchestrator decides which model to use, what context to inject, how to recover from a bad response, and where to draw the line between automated action and human approval. Strong orchestration is what makes the difference between a demo and a system you can rely on.
How Agile Visa uses this term: orchestration is a deep dive in our AI Engineer track, paired with eval design.
See also: MCP, Multi-Agent Systems
M
MCP (Model Context Protocol)AI
MCP, the Model Context Protocol, is an open protocol that lets AI models connect to tools, data sources, and applications through a standard interface. Before MCP, every integration was bespoke. With MCP, the same connector can serve any compliant model client. Practically, MCP makes it easier for enterprises to plug models into mail, calendars, CRM, code repos, and internal data systems without writing a new connector each time.
How Agile Visa uses this term: our AI Engineer cohort builds an MCP server as a graded capstone, so learners ship a working integration, not a slide deck.
MCESFunding SG
MCES, the Mid-Career Enhanced Subsidy, is a higher Singapore course-fee subsidy tier for Singaporeans aged 40 and above, for selected SSG-approved courses. The subsidy is applied at point of enrolment, with the learner or employer paying only the net fee. MCES is the lever that turns many AI and Agile programs from a stretch budget item into an everyday investment for mid-career professionals.
How Agile Visa uses this term: for SSG-supported programs, we surface the MCES net fee for SG mid-career learners up front.
Model-Aware BacklogAgile Visa term
A Model-Aware Backlog is a backlog that captures model selection, prompt assets, eval criteria, and fallbacks as first-class items, not implementation footnotes. In a Model-Aware Backlog, a story about an AI feature names which model it will call, what evals must pass, and what happens when the model is unavailable or wrong. The backlog therefore becomes a real plan of record for an AI system, not just a feature wish list.
How Agile Visa uses this term: we coach AI-Augmented Product Owners to refactor their existing backlog into a Model-Aware Backlog before adding new AI work.
See also: AI-Augmented Product Owner, Backlog
Multi-Agent SystemsAI
Multi-Agent Systems are architectures where several specialised AI agents collaborate, route work to each other, and share state to complete tasks. A common pattern is a planner agent that breaks down a goal, several worker agents that handle specific tool calls, and a critic agent that checks output. Multi-agent designs can unlock harder workflows, but they also multiply failure modes. Good designs invest heavily in observability and tight scopes per agent.
How Agile Visa uses this term: in our AI Engineer track, multi-agent design is taught with a strong bias toward small, well-scoped agents.
MyCoIDFunding MY
MyCoID is the Malaysian Companies Commission identifier used across HRD Corp, SSM, and several other agencies to register employers and trainers. For HRD Corp claimable training, the employer needs an active MyCoID-linked HRD Corp account, and trainers must register their profile with the right MyCoID details. Many delayed claims trace back to a mismatch between the company's MyCoID record and the data on the training application.
How Agile Visa uses this term: our MY onboarding checklist verifies MyCoID details before a claimable program is scheduled.
P
Product OwnerAgile
The Product Owner is the Scrum accountability for maximising the value of the product by owning the backlog and outcomes. The role is not project manager, not analyst, and not stakeholder spokesperson. Strong POs hold a clear product thesis, know the customer, and stay close enough to the team to make trade-offs in real time. In AI-Native delivery, the PO also becomes the chief eval designer of the team.
How Agile Visa uses this term: AI-Augmented Product Owner is our target role for incumbent POs adopting AI-Native Agile.
Prompt-driven DiscoveryAgile Visa term
Prompt-driven Discovery is a discovery practice where structured prompts replace much of the manual desk research, interview synthesis, and option mapping work. Teams keep human discovery for the high-judgement parts. Reading the customer's body language. Sensing what is not being said. The structured prompts speed up the rest. The output is a sharper hypothesis, faster, with the time saved reinvested in real customer conversations.
How Agile Visa uses this term: this is one of the first practices an AI-Augmented Product Owner adopts in our cohorts.
Prompt EngineeringAI
Prompt Engineering is the craft of designing inputs that reliably steer a language model toward useful, on-policy outputs. The work is part writing, part testing. A good prompt has clear roles, an explicit task, the right context, and constraints. Strong practitioners do not stop at a clever phrase. They build evals so they can see when a prompt regresses and why.
How Agile Visa uses this term: in our cohorts, prompt engineering is taught alongside eval design, never on its own.
See also: RAG, Fine-tuning
PSEAFunding SG
PSEA, the Post-Secondary Education Account, is a Singapore savings account that can be used to pay for approved post-secondary education and training expenses. PSEA funds can flow to a wide set of approved providers and programs, including selected SkillsFuture-approved courses. The account is most useful for younger learners and family members of a PSEA holder, since balances can be used across siblings under set rules.
How Agile Visa uses this term: we surface PSEA eligibility on SG program pages where the course qualifies.
R
RAG (Retrieval-Augmented Generation)AI
RAG, Retrieval-Augmented Generation, is a pattern where a model retrieves relevant documents from a knowledge store, then conditions its response on that retrieved context. The pattern lets a general-purpose model answer questions about private or fresh information without retraining. RAG quality depends as much on the retriever as on the model. Most production RAG failures trace back to bad chunking, weak embeddings, or missing source filters, not to the language model itself.
How Agile Visa uses this term: in our AI Engineer cohort, RAG is the first production pattern learners ship before they touch agents.
See also: Fine-tuning, LLM Orchestration
Responsible AIAI
Responsible AI is an operating practice that bakes fairness, transparency, accountability, and human oversight into every AI use case. It is broader than AI Safety, and broader than AI Governance. It is the day-to-day judgement that teams apply when they decide what to ship, what to log, what to disclose, and what to refuse. Done well, Responsible AI is a competitive advantage, since customers trust products that behave predictably.
How Agile Visa uses this term: every cohort closes with a Responsible AI checklist tailored to the learner's organisation.
S
SAFeAgile
SAFe, the Scaled Agile Framework, is a body of patterns for coordinating many Agile teams under shared objectives. It defines roles, cadences, and artefacts at the program and portfolio levels, with the goal of aligning hundreds or thousands of practitioners. SAFe is widely used in regulated enterprises and government. Its critics argue that it can over-formalise. Its supporters argue that it gives large organisations a common language to start from.
How Agile Visa uses this term: we cover SAFe pragmatically in our enterprise programs, while teaching leaders when scaling patterns help and when they hurt.
ScrumAgile
Scrum is a lightweight Agile framework where a cross-functional team delivers value in time-boxed sprints. Scrum defines three accountabilities, Product Owner, Scrum Master, and Developers, five events, and three artefacts. The framework is intentionally minimal. It does not prescribe how to engineer, design, or scale. It asks the team to inspect and adapt their own way of working, sprint by sprint.
How Agile Visa uses this term: Scrum is the baseline framework our AI-Native Agile programs extend, not replace.
Scrum MasterAgile
The Scrum Master is the Scrum accountability for the effectiveness of the team, coaching the system around it to remove impediments. The Scrum Master serves the team, the Product Owner, and the wider organisation. Strong Scrum Masters build a coaching toolkit, watch the team's working agreements, and steer the team away from process theatre. The role is full-time work, not a part-time chair for stand-ups.
How Agile Visa uses this term: most of our Agile learners enter through the Scrum Master pathway and progress from there.
SFECFunding SG
SFEC, the SkillsFuture Enterprise Credit, is a credit that helps eligible Singapore employers offset workforce transformation costs, including selected training. SFEC is meant to encourage employers to invest in their own people. Eligibility is based on specific levy and headcount conditions, and the credit can be used across a defined window. Employers planning AI and Agile uplift programs often combine SFEC with course-level subsidies.
How Agile Visa uses this term: we map SFEC eligibility into the funding section of every SG enterprise proposal.
SkillsFuture CreditFunding SG
SkillsFuture Credit is a Singapore government credit that citizens aged 25 and above can use to offset fees for approved courses. The credit is paid directly to the training provider on the learner's behalf, reducing the out-of-pocket fee. SkillsFuture is one of the most used funding instruments for individual learners in Singapore, particularly for mid-career professionals exploring AI and Agile pathways.
How Agile Visa uses this term: SkillsFuture Credit is accepted on selected SG programs and is highlighted on each eligible course page.
See also: MCES, PSEA
Solutions ArchitectRole
A Solutions Architect designs end-to-end solutions across systems, data, and integrations to meet a defined business outcome. The role bridges engineering and business stakeholders, makes trade-offs visible, and protects long-run maintainability. AI is reshaping the role. A modern Solutions Architect now also designs how models, agents, retrieval, and orchestration sit inside the wider system.
How Agile Visa uses this term: Solutions Architects in our enterprise engagements are central partners when we wire AI into existing delivery.
SprintAgile
A sprint is a fixed-length iteration, usually one to four weeks, that produces a usable increment of product. The sprint is the heartbeat of a Scrum team. Length stays stable. Scope is committed to inside the box. Outside the box, leadership can change direction freely between sprints. A team that keeps sprint length steady gains the right to make sharper, faster decisions over time.
How Agile Visa uses this term: in AI-Native Agile, the sprint becomes an Augmented Sprint, with the same time-box and new work mix.
See also: Augmented Sprint
Story PointsAgile
Story points are a relative-size estimate of effort, complexity, and uncertainty assigned to a backlog item. The unit is intentionally abstract. Two-point items are about twice the size of one-point items, and the team calibrates over time. Story points fail when they get reused as performance targets across teams, since the underlying scale is local and not portable.
How Agile Visa uses this term: with AI agents in the team, we coach teams to split story points from agent work units and report Hybrid Velocity.
See also: Hybrid Velocity, Velocity
U
UTAPFunding SG
UTAP, the Union Training Assistance Programme, is an NTUC subsidy for union members who take up approved training. UTAP reimburses a share of the course fee after completion, up to a yearly cap, with higher caps for older members. For Singapore mid-career professionals who are NTUC members, UTAP often stacks with SkillsFuture Credit to bring the net fee down further.
How Agile Visa uses this term: UTAP eligibility is flagged on each SG course page where the program qualifies.
V
VelocityAgile
Velocity is the amount of backlog a team typically completes per sprint, used as a planning aid. Velocity is a forecasting input. It is not a productivity score. Two healthy teams can have very different velocities and both be excellent. Velocity numbers go wrong fast when leaders treat them as targets, since teams then learn to inflate estimates instead of telling the truth.
How Agile Visa uses this term: classical velocity sits alongside Hybrid Velocity once a team starts using AI agents as part of delivery.