Planning Document
Under review

AI Solutions team plan

A first-year delivery plan for three regional hires using AI Marketplace and Requirements Agent to turn service-line demand into governed, measurable AI solutions.

AI Solutions team: Meg Patakota, with accepted offers from Divya Parekh and Saravan Kumar joining end of August, and Ansh Bahuguna joining mid-October.

At a glance

This plan explains how three AI Solutions hires will help service teams move from AI interest to useful internal tools. The team will use AI Marketplace and Requirements Agent to find high-value opportunities, prioritise them, build rapid prototypes, check governance, and measure whether the work saves time or improves quality.

Purpose

What this team is for

Sets the team up as a delivery bridge between service-line needs, governed prototypes, and measurable improvement.

AI Marketplace Live delivery Reusable accelerators

Embed AI into real consulting delivery, not side experiments

The AI Solutions team should turn repeatable consulting challenges into AI-enabled workflows, micro-apps, automations, and reusable assets that teams actually use. AI Marketplace is the delivery surface for routing demand, validating outputs with users, and packaging the strongest examples for reuse.

Build working solutions Move quickly from unclear business problem to tested prototype or workflow. Reduce manual effort Target document, data, reporting, and process work where time saving can be shown. Scale repeatable assets Turn useful examples into templates, prompts, checks, and app patterns. Enable through doing Coach teams in the context of live work so adoption is practical, not theoretical.

Operating Cycle

All components run together

Shows the shared method every hire should use; the same cycle applies, even when service-line priorities differ.

1Assess maturityWorkflows, pain points, data, sponsors, users, and governance needs.
2Triage demandRequirements Agent captures the problem, gaps, and likely route.
3PrioritiseImpact, feasibility, cost recovery, reuse potential, risk, sponsor readiness, and capacity.
4PrototypeUse Claude Code, AI APIs, low-code platforms, and existing patterns where appropriate.
5Enable by doingWork alongside colleagues on real problems and help them build confidence.
6Package reuseTurn useful examples into tools, prompts, templates, checks, and playbooks.
7Measure impactTrack adoption, time saved, quality, reuse, and scale or stop decisions.

How Requirements Agent supports the cycle

Intake Triage Visibility

Turn an AI idea into a usable case record

Requirements Agent captures enough information to decide the next route before detailed scoping. It is not approval to build; it makes the case visible, comparable, and easier to prioritise.

1. Capture the caseProblem, owner, users, data, output, success measure, and missing information.
2. Route the workMarketplace prototype, existing tool, AI Productivity / L&D, Power Platform, reporting, OCR, custom build, or not ready for build yet.
3. Show the queueLeader visibility of requests, prototypes, usage, blockers, duplicates, and priority cases.

Governance check for Marketplace micro-apps

When to reviewAny internal micro-app, AI workflow, or prototype proposed for AI Marketplace build, registration, reuse, or client-work support.
What to captureRequirements Agent should capture the owner, users, problem, workflow, data, intended outputs, client-work use, access needs, success measure, and any existing demo or walkthrough.
What to checkData sensitivity, client confidentiality, approved platform or model route, access controls, output quality, ownership, support path, and responsible-AI risks.
DecisionConfirm whether the micro-app is ok to build or use, needs improvement, should be registered, should be retired, or should move into a governed Marketplace build with IT and Security input.

Delivery Model

Regional service work plus platform delivery

Shows the initial service-line focus for each candidate while the team continues to share Marketplace delivery responsibilities.

These are initial relationship and delivery priorities, not permanent ownership boundaries.

Candidate service-line focus

Candidate Primary focus Interim coverage
UK candidate SIR · SSO CSCC
US candidate CPD · M&A SPSC
India candidate SPSC · CSCC —

Shared platform-level delivery

Platform responsibility What the three hires need to do Why it matters
AI Marketplace maintenance Keep app pages, reusable assets, guidance, backlog hygiene, and handover routes current. The Marketplace needs to stay useful after prototypes launch; otherwise good work becomes hard to find or maintain.
Requirements Agent operations Review queued cases, improve triage fields, support alerts, and help create GSL/RSL visibility of demand and usage. Leaders need to see what teams are asking for, what has been built, what should be improved, and what should be prioritised next.
AI security and governance review Ask teams to share access or walkthroughs for AI tools built outside Marketplace, especially before client-work use, and work closely with IT and Security to vet them. The team needs to spot data, security, responsible-AI, ownership, support, and quality risks before informal tools become business critical or are used on client work.
L&D and AI Productivity engagement Feed practical examples into learning material, support targeted sessions, and align on when needs are training versus prototype delivery. Adoption should not depend only on the AI Solutions team; learning routes need real examples from service work.
Roadshows and communications Run regional roadshows, share demos, publish short updates, and tell clear stories about impact and reuse. People need to understand what is available, what is possible, and how to bring a good use case forward.

Timeline

Staggered starts, fast ramp

Sets out the ramp from completed setup to delivery: one SIR, one CSCC, and one SSO app or prototype by mid-October, then three more across the six-service-line portfolio by December.

Phase 1 · Now to end August

Set the operating system

Use Meg's current capacity to make the team ready to move fast as soon as the first two hires start.

  • Set up one Kanban board with business and technical swimlanes.
  • Confirm one SIR, one CSCC, and one SSO workflow shortlist.
  • Capture user stories for the SSO consultant-built app already on AI Marketplace.
  • Prepare onboarding packs for Marketplace standards, Claude Code, repo templates, and evidence capture.
Completed Live backlog, governance checklist, evidence template, onboarding pack, one SIR, one CSCC, and one SSO workflow shortlisted, and SSO user stories captured.
Phase 2 · End August to mid-October

Deliver first three outputs

Divya and Saravan join, work with Meg across SIR, CSCC, and SSO, and create the first reusable delivery patterns.

  • Onboard through live Requirements Agent reviews and Marketplace build walkthroughs.
  • Agree scoring for impact, feasibility, cost recovery, reuse, risk, and sponsor readiness.
  • Configure Requirements Agent fields, alerts, and GSL/RSL visibility.
  • Run maturity discovery across SIR, CSCC, and SSO with sponsors and named users.
  • Identify at least six good-value app candidates across the three service lines.
  • Deliver one app or prototype each for SIR, CSCC, and SSO.
  • Use weekly demos to validate outputs and decide what improves, scales, or stops.
  • Capture first reusable prompt, validation, app, and feedback patterns.
Visible by end One SIR, one CSCC, and one SSO app or prototype, six prioritised app candidates, agreed scoring, Requirements Agent visibility, first savings baselines, and a reusable build pattern.
Phase 3 · Mid-October to December

Extend to six outputs

Ansh joins, the CPD track strengthens, and the team expands from the first three service-line outputs into the wider portfolio.

  • Improve the Phase 2 SIR, CSCC, and SSO outputs using user feedback and usage evidence.
  • Deliver three additional apps or prototypes across the six service lines: CPD, M&A, SSO, CSCC, SIR, and SPSC.
  • Prioritise the extra three by impact, feasibility, cost recovery, reuse potential, and sponsor readiness.
  • Complete governance checks before wider reuse or client-work support.
  • Run targeted demos with L&D and AI Productivity.
  • Publish first Requirements Agent visibility view or agreed spec.
  • Record time saved, quality, cost recovery, and scale/improve/stop decision.
Visible by end Six total service-line apps or prototypes, governance record, user feedback, time-saving evidence, Requirements Agent visibility progress, and next decision for each output.
Phase 4 · January onward

Improve, reuse, and scale

Use the first outputs to mature service-line demand, strengthen the Marketplace, and choose the next portfolio with evidence.

  • Improve or retire outputs based on usage, support effort, governance, and value.
  • Convert strong prototypes into reusable Marketplace assets.
  • Package prompts, templates, checks, code patterns, examples, and playbooks.
  • Use Requirements Agent demand to choose the next portfolio.
  • Run a focused hackathon or showcase using real service-line problems.
  • Review measures monthly with service and regional leaders.
Visible by end Reusable asset library, named superuser network, adoption dashboard, focused hackathon, impact story, and next portfolio plan.

Measures And Next Steps

Evidence, not activity

Defines the evidence leaders should see regularly and the immediate actions needed to start managed delivery.

DemandRequirements Agent requests by service line, region, route, priority, and status.Decides what to build next and what should be redirected, merged, or stopped.
DeliveryPrototypes started, validated, improved, scaled, handed over, or retired.Shows whether the team is moving from ideas to usable outputs quickly enough.
AdoptionActive users, repeat users, named superusers, roadshow follow-ups, and reuse of assets.Shows whether tools and materials are being used beyond the original request team.
ImpactTime saved, cycle-time reduction, quality improvement, avoided manual effort, and cost recovery where relevant.Supports improve, scale, stop, or fund decisions.
GovernanceTools reviewed, issues found, IT/Security reviews completed, and client-work usage decisions.Shows whether informal AI use is being made visible and managed responsibly.

Immediate next steps

  • Confirm sponsors and test users for SIR, CSCC, and SSO.
  • Agree maturity assessment dimensions and roadmap format.
  • Finalise Requirements Agent triage categories, alerts, and GSL/RSL visibility fields.
  • Use the Kanban board to manage build tasks, business tasks, blockers, validation, and impact evidence.
  • Review queued cases against impact, feasibility, cost recovery, and reuse potential.

Budget And Enablers

Current-year ask

Captures the two separate TXL funding requests needed to keep the Marketplace running and give the AI team rapid prototyping capacity for the remainder of FY27.

TXL approval in review CDE Automation / Tech budget
Marketplace operating and development envelope
  • Amount: $10,000 capped request for August to December 2026.
  • Covers: AWS hosting, production AI-model usage, and development LLM usage for AI Marketplace.
  • Breakdown: $6,825 core operating allowance, including 5% reserve; $3,175 development allowance.
  • Development use: Repeated model runs to build, test, debug, compare outputs, and improve tools before release.
External AI accounts for rapid prototyping
  • Amount: $3,080 request for named AI Translator / AI Solutions resources.
  • Covers: External AI accounts for research, prompt testing, workflow testing, output comparison, and early prototypes.
  • Calculation: 3 accounts for September-October and 4 accounts for November-December at $200 per account per month.
  • Reserve: 10% for final pricing, tax, or foreign exchange differences.
Controls, scope, and review trigger
  • Assumption: Targeted growth of approximately 5-25 active users across current production and development services.
  • Cost control: Planning cap only; unused budget should not be spent.
  • Review trigger: More than 25 active users, many new production tools, or broad rollout.
  • Not included: Salaries, recruitment costs, and other non-platform costs.

Appendix

Reference terms

Keeps acronyms available for first-time readers without crowding the main story.

Terms used
  • CPD Capital Project Delivery
  • M&A Mergers & Acquisitions
  • SSO Safe & Sustainable Operations
  • SIR Site Investigation & Remediation
  • CSCC Corporate Sustainability & Climate Change
  • SPSC Sustainable Product & Supply Chain
  • AI Marketplace Internal platform for reusable AI tools
  • Requirements Agent Intake and triage tool for AI ideas
  • GSL / RSL Global / Regional Service Leader
  • L&D Learning & Development
  • CDE Client Delivery Evolution
  • TXL Funding approval route