Three years ago, sustainability and AI ethics sat on the last slide of a CRM evaluation, right before "questions." That has flipped. Procurement teams now ask about emissions reporting, AI audit trails, and data residency in the first call, and the answers decide who stays on the shortlist.
We work across India and East Africa, with customers who sell into Europe, so we see these questions from several regulatory angles at once. Here is our read on what has actually changed and what is still mostly marketing.
Why this moved up the list
Two pressures arrived together.
Disclosure got real. Large companies face mandatory sustainability reporting in more jurisdictions, and they push that requirement down their supply chain. If you sell to them, you are asked for numbers, not intentions. The CRM, which already holds suppliers, accounts, and activity, becomes an obvious place to keep that data.
AI made trust concrete. When a CRM only stored data, "trust" meant uptime and access control. When it starts sending customer data to models and letting agents act, buyers want to know exactly what leaves the platform, what is retained, and who is accountable when an agent gets it wrong.
ESG dashboards: what is real
Salesforce's answer on the sustainability side is Net Zero Cloud: emissions tracking across scopes, energy and supplier data, and report templates aimed at common disclosure frameworks.
The honest version: the dashboard is the easy part. The hard part is the data feeding it. Scope 3 emissions in particular depend on supplier figures that are estimated, late, or missing. A beautiful dashboard over guessed inputs is a liability, because it looks authoritative.
A carbon dashboard is only as honest as the spreadsheet a supplier filled in last quarter.
What we would actually do: start with the data you control (your own energy, travel, facilities), document every estimation method next to the number, and treat supplier data as a collection workflow in the CRM, with owners, due dates, and evidence attached, rather than a figure someone types in once a year.
Ethical AI governance: what to actually check
Every vendor now has a trust page. The useful question is which controls you can switch on, inspect, and prove. On Salesforce, that conversation centres on the Einstein Trust Layer and Agentforce guardrails. Whatever the platform, this is our checklist:
- What leaves the platform. Is sensitive data masked before it reaches a model? Is there a zero-retention agreement with the model provider?
- Audit trail. Can you see the prompt, the grounding data, the response, and the action taken, for a specific record, months later?
- Agent boundaries. Agents should run with scoped permissions, not as an all-powerful integration user. Irreversible actions should require a human confirm.
- Output checks. Toxicity and policy filters exist; ask how they are tuned and whether you can review what they blocked.
- A named owner. Someone in your organisation who can pause an agent and explain its decisions. This is a role, not a feature.
If a control cannot be demonstrated in your own sandbox, assume it does not exist yet.
Ask us about governed CRM and AI rollouts→
Privacy is shaping platform choice
For many of the teams we talk to, residency is the first filter, not the last. Where is the data stored? Where is AI inference processed? Can personal data of a Kenyan or Indian customer stay in a permitted region? Data protection laws in India, Kenya, Tanzania, and the EU do not line up perfectly, and a CRM that spans them has to respect the strictest relevant rule for each record.
This is also why we design our own products, like Audexity CRM, with tenancy and data location as first-class decisions instead of afterthoughts. Retrofitting residency is painful. Choosing it up front is cheap.
What is theatre
Some of this is still performance. Watch for:
- Badges without controls. A responsible-AI principles page is not a governance feature.
- Offset-heavy net zero claims with little visible reduction work behind them.
- "Your data is never used for training" with no detail on retention, logging, or sub-processors.
- ESG scores with no methodology. If you cannot see how a number was calculated, you cannot defend it to an auditor.
Where we land
Sustainability and trust are no longer soft criteria. They are part of whether a CRM can legally and reputationally do its job. Evaluate them like any other requirement: ask for the control, test it in a sandbox, and write down who owns it. The platforms that make this easy to prove, rather than easy to claim, are the ones worth shortlisting.
Good to know
What does Salesforce offer for ESG reporting?
Salesforce has a sustainability product, Net Zero Cloud, for tracking emissions, energy, and supplier data and producing disclosure-ready reports. It is useful if your ESG data is already structured; it does not create good data where none exists.
What is ethical AI governance in a CRM?
It is the set of controls around how AI uses customer data and what it is allowed to do: data masking and zero-retention for model calls, audit trails of prompts and actions, toxicity and bias checks, permission boundaries for agents, and a human review path for consequential decisions.
How does data privacy affect CRM platform choice?
Data residency rules, consent requirements, and where AI processing happens can rule platforms in or out before features matter. For teams operating across India, East Africa, and the EU, the question of where data is stored and processed is often the first filter.
We engineer ideas into working software.
Tarka · logic · Shila · stone