Automatic Complaint Summarisation and Triage with Agentforce

Our client, one of the largest social housing providers in the UK, with 40,000 homes and over 90,000 tenants was looking for a new complaints management solution that allowed tenant complaints to be received across a range of digital channels including Email, Web, Voice and a variety of other messaging channels. Alscient recommended Salesforce as a replacement for our customer’s existing solution due to its rich omnichannel customer support offering.

Challenge

Problems associated with our customer’s existing solution included:

• Inconsistent and fragmented complaint records Complaint details were spread across many related records making it difficult to maintain a clear, end-to-end view of the issue.
• Repeated requests for the same information Agents had to manually read through long histories, increasing the likelihood that tenants had to repeat details, causing frustration and delays.
• Slow response times due to manual triage and review Complaints required time-consuming manual checks to understand context, severity and next steps before any action could be taken.
• No automatic classification or prioritisation of complaints Cases were manually categorised by agents, leading to misclassification, inconsistent handling and delayed escalation of serious issues.
• Limited ability to assess tenant sentiment or vulnerability Without automated insight, custom dissatisfaction went unnoticed until the complaint escalated.
• Lack of alignment to complaint handling standards Agents had to manually interpret and apply complaint handling guidance, increasing the risk of inconsistent responses and missed obligations.
• Language barriers for non-English-speaking tenants Messages and emails had to be manually translated, slowing resolution and increasing the risk of misunderstanding for diverse tenant communities.
• Delayed identification of repeat or systemic issues Without automated insight across complaint data, recurring problems affecting multiple tenants were harder to detect and address.
• High reliance on agent experience and judgement Complaint quality and outcomes varied depending on individual agent skill, increasing inconsistency in the tenant experience.

Increased likelihood of escalation and dissatisfaction Manual processes, slower responses, and inconsistent handling increased the risk of unresolved complaints and escalation to external bodies.

What we did

Alscient addressed these requirements following a phased approach.

We initially worked with our client to refine the user stories and provided a solution design which explained how the requirements would be met. Following sign-off of the design we then built out the required functionality over a series of two-week delivery sprints. After each sprint a Show & Tell session was held that demonstrated the capability and provided a mechanism to incorporate feedback.

Following build, the solution underwent system testing and then training was provided on usage of the solution. UAT support was then provided to fully ensure the system fully met internal requirements. Following successful completion of UAT, the solution was promoted to the live environment and used by the Complaints Handling team.

Results

The solution was delivered over a twenty-week period and the following key benefits were realised.

Automatic, end-to-end complaint record summarisation
Agentforce provided GenAI-driven record summarisation across Case, Case Comments, and Email Messages, giving a clear, consistent complaint narrative aligned to the Housing Ombudsman Complaint Handling Code requirement for accurate record-keeping.

Code-aligned complaint triage and classification
The Agentforce solution automatically classified complaints by stage, severity, and theme with explicit reference to the Ombudsman Code, supporting correct handling timescales, escalation paths, and outcomes.

Reduced manual review and faster triage decisions
Automated understanding of complaint context reduced the need for agents to manually read and interpret long histories, significantly reducing handling time and reliance on individual judgement.

Improved compliance with Ombudsman response standards
Agentforce highlighted required actions, response deadlines, and escalation triggers defined by the Code, reducing the risk of missed obligations or non-compliant responses.

Consistent handling regardless of agent experience
Agentic reasoning ensured complaints were assessed and progressed consistently, improving fairness and predictability for tenants and reducing variability between agents.

Multilingual tenant communication at scale
Prompt-driven translation enabled inbound and outbound messages to be translated automatically, supporting clear two-way communication with a diverse tenant base without manual effort.

Faster and more accurate tenant responses
Automated summarisation and translation reduced delays caused by manual drafting, interpretation, and checking, improving tenant satisfaction and trust.

Reduced operational overhead and after-case work
Automation across summarisation, classification, and communication significantly lowered the volume of manual checks, freeing agents to focus on resolution rather than administration.

Clear audit trail for internal review and external scrutiny
Structured summaries, code-aligned classifications, and translated communications created a defensible, transparent record that could be used in defense of ombudsman investigations.

Expected Benefits
Common industry metrics suggest that our solution will deliver 30–60% staff time savings by automating triage, classification, and severity assessment. Customer response times are also likely to improve by 40–70%, as complaints are prioritised and routed correctly on first review. Additional further benefits to be expected include higher first-pass resolution rates, reduced escalation volumes, and better audit readiness. We also expect that customers will also benefit from clearer management reporting, earlier identification of systemic issues, and improved customer trust driven by faster, more transparent complaint handling.