A1: Detailed Dimension Assessments

Data Estate & Ownership 2 / 5 DEVELOPING

Clear ownership by Sales Leader of customer data in Salesforce, but very limited visibility into broader data estate. No documented inventory of data assets beyond knowing contact details and order history exist.

Strengths

• Clear data ownership assigned to Sales Leader
• Core customer data centralized in Salesforce

Gaps

• No visibility into broader technology landscape
• No documented data asset inventory
• Limited understanding of data rights and licensing

Data Estate & Ownership scores Level 2 — Developing. The score reflects a business that has meaningful clarity about one part of its data estate but limited visibility beyond it. The Sales Leader has clear ownership of customer data in Salesforce, and for the immediate automation use case, that is the most important starting point. The core customer record — contact details and order history — lives in one place and has a named owner. That is a more functional starting point than many organisations at this scale. The gap is in breadth, not depth. Beyond Salesforce, there is no documented picture of what other data exists in the business. If any customer or order data lives in spreadsheets, email threads, shared drives, or third-party tools, it is invisible to planning. For the order-status automation use case, this matters practically: if Salesforce does not hold the complete and up-to-date order record — for example, if order fulfilment status is updated in a separate system and only periodically synced — the automation will give customers incomplete or lagged information. There is also an unresolved question about data rights. It has not been confirmed whether the organisation's Salesforce licence permits using the data within it for AI model training or automation purposes. This is a routine licence check, not a likely blocker, but it needs to be confirmed before connecting any external tool to the data. Reaching Level 3 requires a simple data asset inventory: a single document listing every place customer or order data is stored, who owns it, and whether it is the system of record or a secondary source. This is a one-time exercise that takes a few hours and removes a significant planning blind spot. It should be completed in parallel with the technology audit, as both involve asking the same question: what systems does the business actually use?

Data Quality & Trust 2 / 5 DEVELOPING

Salesforce data described as 'accurate' but no evidence of formal quality measurement or management processes. Quality assessment based on user perception rather than systematic evaluation.

Strengths

• User confidence in Salesforce data accuracy

Gaps

• No formal quality measurement processes
• No systematic quality monitoring
• No evidence of quality issue identification or remediation

Data Quality & Trust scores Level 2 — Developing. The score reflects user confidence in Salesforce data accuracy without any formal basis for that confidence. The team works with Salesforce day-to-day and believes the data is accurate — that operational familiarity is real and not worthless. But for an AI system, perception of accuracy is not sufficient. Automation makes decisions based on what the data says, not what the team believes it says. The specific risk for the order-status use case is around completeness, consistency, and timeliness of the fields the automation will depend on: customer name, contact method, order reference, and order status. If a meaningful percentage of records have blank contact fields, inconsistent order reference formats, or an order status that is updated manually and therefore lags behind real fulfilment, the automation will either fail to process calls or give customers incorrect information. The latter outcome is worse than the current manual process. No process exists to identify or remediate quality issues systematically. If a data problem is found, there is no documented step for fixing it and preventing recurrence. There is also no data lineage — it is unknown how order status information gets into Salesforce, how often it is updated, or whether it reflects real-time warehouse or logistics data or is entered manually by a team member. The practical fix is a structured quality review of the four or five specific fields the automation will use, conducted before the pilot is built. This does not require specialist tools — a systematic check against a defined standard, documented and acted on, is sufficient to move to Level 3. The ongoing quality habit established afterwards (a monthly check using the same fields and standards) is what sustains that level over time.

Technology & Integration 1 / 5 UNSTRUCTURED

Only Salesforce identified as a system, described as 'simplest implementation.' No knowledge of other systems, integration capabilities, or API availability. Complete gap in technology architecture understanding.

Strengths

• Salesforce as centralized customer system

Gaps

• No visibility into integration architecture
• No knowledge of API capabilities
• No understanding of system interconnections
• Simplest Salesforce implementation suggests limited functionality

Technology & Integration scores Level 1 — Unstructured. This is the most consequential single finding in the assessment and the primary reason the Imbalance Rule has been applied. The score reflects not a weak or developing infrastructure, but an infrastructure whose capabilities are entirely unknown — and a Salesforce implementation described as the simplest possible, which in practice suggests an out-of-the-box configuration with no integrations, no workflow automation, and no confirmed API access. API access is not an optional feature for AI adoption — it is the technical gateway through which any external automation tool communicates with Salesforce. Without it, no tool can read order status data, trigger responses, or update records. The automation use case depends entirely on this capability being present or being enabled. Beyond the API question, no one in the organisation currently has visibility into the broader technology landscape. It is unknown what other software tools the business uses day-to-day, whether any of them hold relevant customer or order data, and whether any of them are connected to or affect the Salesforce record. This creates a planning blind spot: if order fulfilment is managed in a separate system and Salesforce is not receiving real-time updates from it, the data quality problem (D2) and the technology problem (D3) are directly linked. The most urgent action is a technology audit — a single scoping session, ideally with a Salesforce consultant, that answers four specific questions: What Salesforce edition is in use? Does it include API access? What other tools does the business use? What would it cost or require to enable connectivity? This audit is fast, low-cost, and removes the most critical unknown in the entire assessment. Every subsequent technical action on the roadmap depends on its findings.

Governance, Ethics & Compliance 1 / 5 UNSTRUCTURED

Explicitly stated 'no policies' for data access or changes. No evidence of governance structures, accountability frameworks, or compliance considerations.

Gaps

• No formal data access policies
• No governance structures
• No accountability frameworks
• No evidence of compliance considerations

Governance, Ethics & Compliance scores Level 1 — Unstructured. The assessment explicitly confirmed that no policies exist governing data access, data changes, or data handling. There are no documented rules about who can view or edit customer records in Salesforce, no accountability framework for data decisions, and no assessment of the organisation's obligations under data protection regulations including GDPR, which applies to UK businesses handling personal data. For the immediate automation use case, this creates two practical risks. First, connecting an external AI tool to Salesforce without any documented rules about what it is and is not permitted to do means there is no framework for managing the tool's behaviour — if it accesses records it should not, shares information incorrectly, or makes an error affecting a customer, there is no defined process for responding. Second, if a customer asks how their data is being used in an automated system — a routine and legally legitimate question under GDPR — the organisation currently has no documented answer. The good news is that reaching Level 2 does not require a formal governance programme. For an organisation of this scale, three documented rules are sufficient: who is permitted to access customer data and under what circumstances, what an automation tool is and is not permitted to do with that data, and how the organisation will respond if a customer raises a concern. A single page, completed by the Sales Leader, is the target output — not a policy framework. This is deliberately the first action on the roadmap because it is the fastest to complete, requires no technical work, and removes a legal and reputational risk that exists independently of whether AI is ever introduced. It also establishes the habit of documenting data decisions, which will be needed at every subsequent stage of the automation journey.

AI Use Case Readiness 2 / 5 DEVELOPING

Clear use case identified (customer support automation for 10 daily calls) with relevant data in Salesforce. However, current process is entirely manual with team spending significant time updating order status.

Strengths

• Well-defined use case with clear business need
• Relevant customer and order data available in Salesforce
• Quantified volume (10 calls per day)

Gaps

• Entirely manual current process
• No automation infrastructure
• Unknown data structure optimization for AI

AI Use Case Readiness scores Level 2 — Developing. Within the constraints of a Pre-Activation overall score, this dimension represents the clearest strength in the assessment. The organisation has identified a specific, narrow, and commercially relevant automation target: handling the 10 daily calls where customers are asking for their current order status. The volume is quantified, the use case is bounded, and the relevant data already exists in Salesforce. That combination — a defined problem with an accessible data source — is the right starting point for a first AI deployment. The score is Level 2 rather than higher because a well-defined use case is not the same as a ready use case. The current process is entirely manual with no automation layer in place — there is no existing workflow, trigger, or integration that an AI system could be built on top of. The data fields the automation would depend on have not been verified for completeness or structure. And the technical pathway — whether Salesforce can connect to an automation tool at all — remains unconfirmed. This dimension has a direct dependency on three others. Technology & Integration (D3) must be resolved first: until API connectivity is confirmed, the use case cannot be technically delivered regardless of how clearly it is defined. Data Quality (D2) must be addressed second: the automation will only produce accurate outputs if the Salesforce fields it queries are consistently populated and current. And Governance (D4) must be in place before an external tool is given access to live customer data. The sequencing implication is important for expectation-setting: AI Use Case Readiness cannot advance in isolation. It is the destination, not the starting point. The roadmap is structured to build the enabling conditions first and deliver on this dimension's potential last — at which point the foundations will be in place to do it reliably.

Leadership, Culture & Capability 2 / 5 DEVELOPING

Sales Leader driving AI initiative shows leadership commitment, but team explicitly has 'no tech capabilities.' Significant capability gap for AI adoption and operation.

Strengths

• Clear leadership commitment from Sales Leader
• Identified business need driving initiative

Gaps

• Team has no technical capabilities
• No evidence of data literacy
• Significant training requirements for AI adoption

Leadership, Culture & Capability scores Level 2 — Developing. The Sales Leader driving this initiative represents a genuine and important asset. There is a clearly articulated business need, a realistic use case, and the organisational commitment to pursue it. At a small organisation, leadership-driven initiative is often the most important enabler of change — and that ingredient is present here. The gap is on the capability side, and it is significant. The team has no technical skills. There is no internal resource who could evaluate AI tools, configure integrations, manage API connections, or troubleshoot automation failures. There is also no evidence of data literacy beyond the Sales Leader's operational familiarity with Salesforce — the team does not currently have the skills to assess data quality systematically, interpret outputs from an automated system, or identify when automation is producing incorrect results before those errors reach customers. This capability gap does not make the use case unachievable — it makes external support essential. The organisation cannot build this automation internally. The practical response is to identify a Salesforce-experienced implementation partner who can handle the technical build while the Sales Leader makes strategic decisions. That division of responsibility — external partner for technical execution, Sales Leader for business judgement — is a realistic and appropriate model for a first AI deployment at this scale. The Sales Leader's parallel investment in a short AI literacy course serves a specific purpose: it ensures the organisation can evaluate the partner's recommendations with enough understanding to ask the right questions, avoid being oversold, and make informed decisions about what to build and what to defer. It does not require the Sales Leader to become technical — it requires enough working knowledge to stay in control of the process.

A2: Complete Gap Analysis

Data Estate & Ownership

Level 2 → Level 3

You know what customer data you hold in Salesforce and who owns it — that is a genuine strength. But beyond that, there is no documented picture of what other data exists in the business, where it lives, or whether you have the right to use it for AI purposes. For customer support automation, you need to be confident that the order and contact data in Salesforce is complete, that nothing critical sits in spreadsheets or email threads outside the system, and that you understand any licensing or usage constraints on that data. Reaching Level 3 means producing a simple, documented inventory of your data assets — even if that inventory is short — and confirming that the data you plan to use for automation is fully within your control.

Root Causes

• No documented data asset inventory exists beyond an implicit understanding that Salesforce holds contact details and order history — meaning any data held outside Salesforce (e.g. in email, spreadsheets, or third-party tools) is invisible to planning
• The Sales Leader has clear ownership of customer data but no visibility into the broader technology landscape, so it is unknown whether other systems exist that hold relevant data or create gaps in the customer record
• No assessment has been made of data rights or licensing — it is unclear whether the organisation has the right to use its Salesforce data for AI model training or automation purposes

Data Quality & Trust

Level 2 → Level 3

The team believes Salesforce data is accurate, but that belief is based on day-to-day experience rather than any systematic check. For customer support automation, this matters practically: if an AI tool pulls an incorrect order status or outdated contact detail, it will give a customer wrong information — which is worse than the current manual process. Reaching Level 3 means moving from 'we think it's accurate' to 'we have checked it and here is what we found.' This does not require sophisticated tooling — it means defining what good looks like for the fields your automation will rely on (order status, customer name, contact method) and doing a structured review of those fields before you build anything on top of them.

Root Causes

• Data quality is assessed through user perception and operational familiarity rather than any structured review — there is no defined standard for what 'accurate' means for the specific fields that customer support automation would depend on
• No process exists to identify or remediate quality issues — if a data problem is discovered, there is no documented step for fixing it and preventing recurrence
• No data lineage or traceability exists, meaning it is unknown how order status data gets into Salesforce, how frequently it updates, or whether it reflects real-time fulfilment status — all of which are critical for an order-update automation use case

Technology & Integration

Level 1 → Level 3

This is the most significant structural blocker for your automation ambition. Only Salesforce has been identified as a system, and it is described as the simplest possible implementation — meaning it is likely to have limited API capability, restricted integration options, and minimal automation features enabled. There is no understanding of what other systems exist, how they connect, or whether Salesforce can actually send and receive data from an AI tool. Before any automation can be built, you need to know what your Salesforce instance can technically do, whether it can connect to external tools, and what the full technology picture of the business looks like. This is a foundational gap that cannot be skipped — it determines whether your use case is even technically feasible with your current setup.

Root Causes

• The Salesforce implementation is described as the 'simplest implementation,' which typically means out-of-the-box configuration with no API integrations, no workflow automation, and no connection to external systems — the minimum viable starting point for AI integration is not in place
• No one in the organisation currently has visibility into the broader technology architecture — it is unknown what other tools, platforms, or systems exist alongside Salesforce, making it impossible to assess integration feasibility
• No knowledge of API availability or capability exists, meaning it is unknown whether Salesforce can be connected to an AI or automation tool without significant configuration or upgrade work

Governance, Ethics & Compliance

Level 1 → Level 2

There are currently no policies governing who can access data, who can make changes to it, or how it should be handled — this was explicitly confirmed during the assessment. For a small organisation automating customer support, this creates two practical risks: first, if an AI tool is given access to Salesforce, there is no framework for deciding what it should and should not be able to do; second, if a customer asks how their data is being used, there is no documented answer. Reaching Level 2 is the realistic and sufficient target here — you do not need a full governance programme, but you do need a small number of documented rules: who can access customer data, what an AI tool is permitted to do with it, and how you would respond if something went wrong. This is achievable quickly and is a prerequisite for responsible automation.

Root Causes

• Explicitly confirmed during assessment that no data access policies exist — there are no documented rules about who can view, edit, or export customer data from Salesforce
• No accountability framework exists for data decisions — if a data error occurs or a customer raises a concern about how their data was used, there is no defined owner or process for responding
• No compliance considerations have been assessed — it is unknown whether the organisation has obligations under data protection regulations (e.g. GDPR or equivalent) that would affect how customer data can be used in an automated system

AI Use Case Readiness

Level 2 → Level 3

Your use case is one of the clearest and most practical in this assessment: 10 order-status calls per day, handled manually, with the relevant data already in Salesforce. That clarity is a real asset. The gap is not in the ambition — it is in the infrastructure needed to execute it. The current process is entirely manual, there is no automation layer in place, and it is unknown whether the Salesforce data is structured in a way that an AI tool could reliably use. Reaching Level 3 means confirming that the data fields your automation would rely on are consistently populated and structured, and that a technical pathway exists to connect Salesforce to an automation tool. This dimension cannot move forward until Technology & Integration (D3) and Data Quality (D2) are addressed first.

Root Causes

• The current order-status update process is entirely manual with no automation infrastructure in place — there is no existing workflow, trigger, or integration that an AI layer could be built on top of
• It is unknown whether the Salesforce data fields relevant to the use case (order status, customer contact details) are consistently structured and populated in a way that an AI tool could reliably query and act on
• Technology & Integration (Level 1) is a direct prerequisite blocker — until the Salesforce API capability and integration architecture are understood, it is not possible to determine whether the use case is technically deliverable with the current setup

Leadership, Culture & Capability

Level 2 → Level 3

The Sales Leader is driving this initiative with clear commitment and a well-defined business need — that is the foundation this whole effort rests on, and it should not be underestimated. The gap is on the capability side: the team has no technical skills, no data literacy, and no experience operating automated systems. This does not mean the use case is unachievable — it means the organisation needs to be realistic about what it can build and manage internally versus what it needs external support for. Reaching Level 3 means the Sales Leader and team have enough working knowledge to make informed decisions about AI tools, evaluate vendor options, and manage a simple automation once it is live — even if the technical build is done by an external partner.

Root Causes

• The team explicitly has no technical capabilities — there is no internal resource who could evaluate AI tools, configure integrations, or troubleshoot automation failures without external support
• No evidence of data literacy exists beyond the Sales Leader's operational familiarity with Salesforce — the team does not currently have the skills to assess data quality, interpret outputs from an AI system, or identify when automation is producing incorrect results
• The organisation has no experience managing automated systems, meaning there is no internal knowledge of what ongoing maintenance, monitoring, or error-handling looks like in practice — creating a dependency risk if external support is not planned for

A3: Detailed Roadmap

A4: Methodology & Scoring

This assessment was conducted using the ARIA six-dimension framework, which evaluates AI readiness across Data Estate & Ownership, Data Quality & Trust, Technology & Integration, Governance, Ethics & Compliance, AI Use Case Readiness, and Leadership, Culture & Capability. Each dimension is scored on a 1–5 maturity scale and weighted according to the organisation's specific AI ambitions, producing a weighted Index Score on a 1.0–5.0 scale. Where any single dimension scores Level 1 (Unstructured), the Imbalance Rule caps the overall Index Band to reflect the structural risk that an aggregate score would otherwise mask — ensuring the report surfaces critical blockers rather than averaging them away.

Dimension Weighting

Dimension	Weight
Data Estate & Ownership	20.0%
Data Quality & Trust	15.0%
Technology & Integration	25.0%
Governance, Ethics & Compliance	10.0%
AI Use Case Readiness	15.0%
Leadership, Culture & Capability	15.0%

Imbalance Rule

The imbalance rule was applied to this assessment. When a structural weakness exists in one or more dimensions, the overall index band is capped to reflect that systemic risk.

• Technology & Integration: Score of 1 (Unstructured) in Technology & Integration represents a structural risk that caps the overall band.
• Governance, Ethics & Compliance: Score of 1 (Unstructured) in Governance, Ethics & Compliance represents a structural risk that caps the overall band.

Scoring Scale

Level	Description
1	Structural Gap: No meaningful capability or infrastructure in place. Fundamental building blocks are absent.
2	Developing: Initial steps taken but capability is fragmented, inconsistent, or ad hoc.
3	Established: Functional capability exists with some standardisation, though gaps remain in maturity or coverage.
4	Advanced: Strong, systematic capability with clear governance, broad adoption, and measurable outcomes.
5	Optimised: Industry-leading capability that is continuously refined, fully embedded, and drives competitive advantage.

A5: Engagement Summary

Organisation	M Plc
Organisation Type	corporate_enterprise
Organisation Scale	small_organisation
Assessment Purpose	internal_transformation_planning
Report Register	functional
Primary Reader	Sales Leader driving customer support AI initiative, needs practical recommendations grounded in operational reality
Engagement ID	804f39e56b24
Report Date	30 April 2026