product leadership
WHAT DOES AN AI-AUGMENTED PRODUCT ORGANISATION LOOK LIKE?
Product teams are surrounded by evidence but still make decisions from periodic summaries, partial memories and whoever argues best. The change that matters is that customer evidence, product behaviour and commercial context become a continuous operating layer around every product decision. The roadmap stops being a presentation of intentions. It becomes a living record of bets, evidence and choices.
Evidence arrives faster than
decisions can absorb it.
A mid-market product function usually contains 3 to 30 core product staff in companies with 100 to 2,000 employees, an estimate consistent with the team-size bands in ProductPlan's 2025 survey and current software-team ratios. The smallest version is a Head of Product with one or two product managers. Larger versions add product directors, portfolio or category managers, product operations, researchers, designers, product analysts and pricing responsibility. Engineering, industrial design, testing and technical lifecycle management may sit beneath a wider product umbrella without belonging to product management itself.
Product sits between market demand and the resources used to develop products. It drives revenue through new offers, adoption, retention, pricing and portfolio mix. It shapes development choices, product complexity, support demand and decisions to stop weak products.
Eight processes- 01Discovery and customer research
- 02Market and competitor intelligence
- 03Roadmap and prioritisation
- 04Requirements and product specifications
- 05Pricing and packaging
- 06Launch, documentation and enablement
- 07Product analytics and feedback loops
- 08Portfolio management
Where the information lives is the problem. Jira, Azure DevOps or Linear. Productboard, Aha!, spreadsheets and presentation roadmaps. Amplitude, Mixpanel or web analytics. CRM, support tickets, call recordings, ERP, PLM, Confluence, SharePoint and the memories of salespeople, researchers and senior product managers. Each system contains part of the answer. Few decisions receive the whole answer in time.
The roadmap often reflects which leaders negotiated most effectively because the evidence could not be assembled at decision speed.
Product teams already track revenue, margin, adoption, activation, retention, churn, satisfaction, time to market, experiment velocity and portfolio return. The constraint is not a lack of signals. It is the work required to connect signals to a decision while the decision is still open.
The roadmap becomes a
living decision record.
The function changes at its centre in two ways.
Customer and market evidence becomes continuous rather than episodic. Approved interviews, tickets, sales calls, reviews and behavioural events become a product memory that can be searched, compared and refreshed. Every important finding carries its source, segment, product, date and validity. Each roadmap bet can show the customer evidence, product behaviour, commercial value, delivery demand and assumptions beneath it. The team sees contradictions and gaps instead of hiding them inside a score.
Artifact production stops setting the product cadence. Briefs, product requirements, release notes, comparisons and portfolio packs can be prepared within the flow of work. Product managers become accountable editors of evidence and choices rather than producers of documents. Their scarce contribution moves toward problem selection, study design, causal reasoning, positioning, tradeoffs and customer commitments.
The function moves from a quarterly reconstruction of customer needs and product performance to a weekly view of what changed, why it matters and which decision now deserves attention.
Where it is weakestMore evidence does not remove strategic judgment. Weak coverage can make the loudest customers look representative. Poor metric definitions can turn correlation into a confident story. Faster analysis makes a weak assumption travel faster unless a person tests the source, the segment and the claim.
A day, a week,
a month.
A product function that invests heavily still employs product managers, researchers and analysts in 2028. Its shape changes. Product operations moves from calendar and template administration toward taxonomy, permissions, workflow evaluation and product-memory quality. Analysts spend less time answering recurring requests and more time on causal analysis. Researchers conduct more studies and test whether synthesis suppresses minority views. Product managers spend more of the week with customers. Product leadership concentrates on product thesis, pricing and portfolio choices.
The dayDiscovery workflows process approved interviews, support conversations, reviews and sales calls. Each product manager receives a short brief covering new needs, contradictory evidence, unusual customer behaviour and competitor moves. Sources are attached. The product manager decides what deserves investigation and which customers to call.
A live workspace brings together the relevant research, usage, commercial exposure, support history and earlier decisions. It prepares an interview guide, an analysis path or a first specification from approved material. The product manager checks coverage, challenges the interpretation and supplies context the systems do not hold.
Approved conversations become cited findings. Decisions record the alternatives considered, the assumptions accepted and the evidence that would reverse them. Draft specifications, internal notes and feedback tags are proposed in the systems that own them. The responsible person approves consequential changes.
The weekEach week, product squads examine active bets through one shared view. It combines qualitative feedback, cohort behaviour, commercial exposure, experiment results and delivery status. Competitor changes are interpreted beside actual win-loss evidence. The meeting starts with contradictions, weak assumptions, unusual movement and choices that need intervention, not a tour through every status line.
The monthProduct leadership reviews the portfolio. The pack shows where usage, margin, retention, support burden and strategic fit are improving or deteriorating. Pricing scenarios and packaging hypotheses are ready for discussion. Leaders decide which experiments to run, which commitments to defer, and which products to invest in, hold, harvest or stop.
Continuous discovery, traceable choices and current portfolio evidence used to be separate ambitions. They become one operating rhythm, with the product manager accountable for what the evidence means.
What runs, and what
stays with the person.
| Process | What the agent does | What stays with the person |
|---|---|---|
| Discovery and customer research | Prepares interview guides from existing evidence, transcribes approved sessions, proposes themes and links each finding to source material | Researchers control sampling and interpretation; product managers conduct important conversations and investigate outliers |
| Market and competitor intelligence | Monitors approved public sources and internal win-loss material, explains what changed and produces a cited brief | Product leadership judges strategic significance and decides whether the change matters |
| Roadmap and prioritisation | Assembles alternative rankings, missing evidence, dependencies, strategic fit and rejected assumptions | The product manager owns the recommendation; product leadership owns capital allocation |
| Requirements and product specifications | Drafts problem statements, product requirements, stories, edge cases and open questions from approved discovery material | Product, design, legal and engineering review the scope, tradeoffs and readiness of the result |
| Pricing and packaging | Prepares segment economics, competitor comparisons, packaging alternatives, interview guides and experiment designs | Product, finance and commercial leaders approve the hypothesis, guardrails and any change |
| Launch, documentation and enablement | Produces audience-specific drafts from the approved specification and decision record, then checks them against release status | Owners approve positioning, claims, readiness, publication and customer commitments |
| Product analytics and feedback loops | Answers governed questions, creates analyses, flags anomalies and proposes explanations with underlying data | Analysts validate taxonomy, causal claims and experiment design; product managers decide what action follows |
| Portfolio management | Maintains invest, hold, harvest and stop scenarios, and exposes changed assumptions, cannibalisation and unsupported investment | Executives decide where resources move and which products continue |
Four stages, and the
product stack today.
GRAIL's Access Ladder has four stages. Each earns the next.
Research, specifications, decisions, policies and launch material form the first product memory. Product managers inspect sources and taxonomy.
Roadmap decisions and analytics compare plans with behavior. CRM, support and conversations deepen discovery. Outputs require evaluation.
The agent proposes updates with sources and affected records. Prices, roadmap commitments, specifications, external claims and experiments with identifiable people require named approval.
Summaries, tags, status updates and monitoring run within limits. The workflow flags anomalies but cannot allocate capital, set prices or commit the roadmap.
Six things we believe,
from building this.
Product functions do not need another layer of generated documents. They need a better relationship between evidence and choice. Six beliefs shape how we build that relationship.
A fast specification is useful, but it is not the structural change. The real asset is an evidence base that preserves source, segment, product, date, validity and decision history so the next choice starts from what the company already knows.
Combining interviews, tickets, sales calls, reviews and behaviour can reveal patterns no single channel holds. It can also turn a noisy sample into a confident answer. The team must know who is missing, which segment is speaking and whether observed behaviour supports the stated need.
Each bet should carry its evidence, commercial exposure, delivery demand and stopping assumptions. That makes political requests easier to examine and weak investments easier to interrupt. The value is not prettier prioritisation. It is a more explicit choice.
Product requirements, release notes and portfolio packs can be prepared from approved evidence and kept current across systems. Product managers should spend the capacity that releases on customers, hypotheses, tradeoffs and commitments. A fuller document does not rescue a poorly chosen problem.
Agents can retrieve evidence, compare histories, expose contradictions and prepare alternatives. People retain study design, causal reasoning, strategic significance, pricing judgment and portfolio choice. Approval must mean inspecting the evidence and accepting the decision, not clicking through a queue.
A delivery system is attractive because its fields are structured and its connection is straightforward. Connected first, it can make briefs optimise activity rather than product outcomes. The chain should run from customer evidence to behaviour, decision and delivery, with each link traceable.
These six beliefs do not favour a particular product tool. They concern how a company learns, chooses and commits. Every connected system inherits them.
Roles, rhythm, and
where it fails.
Product managers move from document ownership toward evidence quality, hypothesis design and decisions. They learn to inspect sources, test segment coverage, challenge causal claims, write clear choices and recognise which actions need approval. Their contribution moves toward problem selection, study design, causal reasoning, positioning, tradeoffs and customer commitments.
Product operations owns taxonomy, permissions, evaluation sets and the quality of the shared product memory. Researchers own sampling, consent and interpretation. Analysts own metric definitions, causal standards and experiment validity. Product leaders define decision rights, protect room for customer contact and keep delivery signals from becoming a proxy for individual performance.
The team needs research design, data literacy, experiment design, source verification, careful handling of customer material and clear decision writing. Prompt familiarity is not enough. People must understand where a synthesis came from, what it excluded and what evidence would change their mind.
The rhythmCurated documents establish the first product memory and working rhythm
A stable weekly rhythm develops around approved evidence
Governed read access connects roadmap, analytics and customer systems
Approved writes and bounded internal actions keep product records current
Where it failsSource coverage is ignored. Loud customers dominate. Themes lose segment context. Sentiment is confused with need. Correlations become causal claims. Stale assumptions remain in product memory. Review work is added without removing the old reporting ritual. Jira is connected before customer evidence, so delivery activity becomes the answer to every question.
The product function changes when the people who make choices also shape the memory, approval boundaries and working rhythm around those choices. The tool is only the carrier.