Alma — AI Tools Proposal

Before you issue an announcement, run the draft through this tool. It compares against the client's recent announcements and relevant peer announcements, flagging tone shifts, language inconsistencies, or sentiment changes.

You catch problems before Primer (or analysts) catch them for you.

Day-to-day workflow:

1. Consultant uploads draft RNS
2. Selects client (system already has their recent announcements stored)
3. Optionally selects peers to compare against
4. Clicks "Analyse"
5. Within 2–3 minutes: receives flagged issues with explanations
6. Makes edits, re-runs if needed

Defensive Value

Inconsistent messaging is the single most common criticism analysts raise about companies. A tone shift between a bullish trading update and a cautious results statement — even when circumstances justify it — creates doubt about management credibility.

This tool catches those inconsistencies before they reach the market. The value isn't in the average case (where drafts are fine), it's in the one case where a subtle tonal shift would have been picked up by Primer or an eagle-eyed analyst.

Competitive Landscape

No financial PR firm we're aware of has automated sentiment screening against historical and peer benchmarks. This gives Alma a process advantage that competitors can't replicate without similar tooling.

Build Cost

Pre-AI estimate: £80,000–150,000 build + £20,000–40,000/year maintenance

Building equivalent capability without AI would mean developing a custom NLP pipeline — sentiment models trained on financial language, a comparison engine for historical announcements, and a scoring framework.

Operational Cost

Without AI, this work falls to senior consultants. At a blended rate of £100–200/hour:

What Wasn't Possible Before

Systematic peer benchmarking on every draft. Manual review compares against memory and intuition. This tool compares against the actual text of recent peer announcements across 7 dimensions simultaneously — something no human reviewer can do in real time, regardless of seniority.

Enter a client's website and their peers. The system crawls the relevant pages, extracts how each company presents itself, and produces a structured audit showing: internal consistency issues, gaps against key messages, and how the client compares to competitors.

Output is a client-ready report.

Day-to-day workflow:

1. Enter target company URL
2. Enter 3–5 peer URLs
3. Input key messages to test for (or use client's existing messaging framework)
4. Click "Generate Audit"
5. Within 10–15 minutes: receive structured report
6. Download as Word/PDF, ready for client presentation

Revenue Opportunity

Alma's most directly monetisable tool. Website messaging audits are a natural upsell for new and existing clients.

The numbers:

• Alma's cost per audit: ~£1.50 (AI costs)
• Charge to clients: £2,000–5,000 per audit
• At just 2–3 audits per quarter: £4,000–15,000/year in additional revenue
• At 1 per client per year (15 clients): £30,000–75,000/year

New Business Weapon

Run a quick draft-mode audit on a prospect's website before a pitch meeting. Walk in with specific, evidence-based observations about their messaging gaps relative to peers. This transforms a generic credentials presentation into a consultative conversation — at essentially zero cost to Alma.

Competitive Landscape

Every IR advisory firm talks about messaging consistency. Few can deliver a structured, peer-benchmarked audit because the manual cost makes it economically impractical at scale. Alma can offer this as standard.

Build Cost

Pre-AI estimate: £120,000–200,000 build + ongoing crawler maintenance

Building equivalent capability without AI would mean developing a custom web crawler, a content extraction engine, a comparison framework, and a report generator.

Operational Cost

A manual website messaging audit is a 2–3 day project for a senior consultant.

What Wasn't Possible Before

Running 20 quick audits per month. Manual audits are too expensive to use speculatively. With AI, Alma can run a draft audit on every prospect before a pitch meeting, on every new client at onboarding, and on every existing client quarterly — turning a rare deliverable into a routine capability.

Produces first drafts of common financial PR documents using client context. Consultants provide context and bullet points, and the AI generates a professional first draft in Alma's house style.

Includes streaming iterative refinement — make edits, adjust tone, or restructure through a conversational interface until the document is right.

Document types:

• RNS announcement drafts
• Investor Q&A briefings
• Holding statements
• Key messages documents
• Peer summary notes
• Board advisory notes

Cost Saving

£48,000–144,000/year in consultant time currently spent drafting, reduced to £1,200/year in AI costs. That's effectively a full-time equivalent redirected from drafting to client strategy.

The numbers:

• 40 documents/month × 2–3 hours each = 80–120 hours/month
• At £50–100/hour blended cost = £4,000–12,000/month freed
• Annual saving: £48,000–144,000 vs £1,200 tool cost

Competitive Landscape

AI-assisted document drafting is becoming table stakes across professional services. The differentiation is in the specificity — calibrated to financial PR document types with Alma's house style, client context auto-loaded, and domain-specific prompt engineering. Competitors using generic AI tools will produce generic output.

Build Cost

Pre-AI estimate: £60,000–100,000 build + ongoing template maintenance

Building equivalent capability without AI would mean developing a template engine with document-type-specific logic, a context management system, and iterative editing workflow.

Operational Cost

Manual drafting at 40 documents per month across 6 document types:

What Wasn't Possible Before

Instant context-aware first drafts. A consultant starting a holding statement currently has to re-read the client's recent announcements, key messages, and history before writing. The AI loads all of this automatically and produces a draft in Alma's house style in minutes — something that would require a dedicated knowledge management system costing far more than the AI approach.

Each morning, consultants receive a client-specific summary of overnight and previous day coverage — their clients, relevant peers, sector news, and analyst commentary.

Written in a 'ready to forward to client' tone, or at least 'ready to scan and decide what matters.'

Currently this process takes around 45 minutes per consultant per day. This tool aims to reduce that to a quick review and forward.

Cost Saving

£90,000–180,000/year in consultant morning time, reduced to ~£4,000/year in tool costs. This is Alma's single largest time drain — 75 hours per month across the team.

The numbers:

• 5 consultants × 45 min/day × 20 days = 75 hours/month
• At £50–100/hour blended cost = £3,750–7,500/month freed
• Annual saving: £45,000–90,000 vs ~£4,000 tool cost
• Every additional client adds £15/month, not another 45 minutes of staff time

Competitive Landscape

Many financial PR firms use Meltwater or Cision (£15,000–30,000/year) which still require manual filtering and summary writing. Alma's summaries arrive ready to forward — a better client experience at lower total cost.

Build Cost

Pre-AI estimate: £100,000–180,000 build + £10,000–20,000/year maintenance

Building equivalent capability without AI would mean developing a news aggregation pipeline, content extraction from multiple sources, and a summarisation/formatting engine.

Alternatively, a managed monitoring service (Meltwater, Cision) costs £15,000–30,000/year — but still requires manual summary writing by consultants.

Operational Cost

What Wasn't Possible Before

Client-ready summaries with zero writing time. Monitoring services aggregate links. Consultants still have to read, filter, and write the summary. This tool delivers a finished, client-ready digest — a capability that requires AI to produce at the quality and frequency needed.

Monitors LSE, ADVFN, Stockopedia, and similar retail investor bulletin boards for client stock discussions. Alerts when sentiment shifts significantly, discussion volume spikes, or concerning topics emerge.

The value is in the one time it catches something important before it escalates — a coordinated short attack, management criticism gaining traction, or takeover speculation building.

Defensive Value

Bulletin board monitoring is insurance. The value is asymmetric — low cost for consistent monitoring, extremely high value when it catches something early.

Retail investor sentiment on bulletin boards can foreshadow institutional and media attention. A thread gaining traction on ADVFN today could become a journalist's story next week. Early warning means Alma can prepare a holding statement, brief the client, or advise on a pre-emptive response before the narrative escalates.

The value is in the one time it catches a coordinated short attack, takeover speculation building momentum, or management criticism gaining traction — before it reaches the press.

Competitive Landscape

Systematic bulletin board monitoring is rare among financial PR firms. Most rely on ad hoc checking or client reports. This gives Alma an early warning system that competitors lack — and that clients increasingly expect as part of comprehensive IR advisory.

Build Cost

Pre-AI estimate: £150,000–250,000 build + £30,000–50,000/year maintenance

Building equivalent capability without AI would mean developing web scrapers for multiple bulletin board platforms, a sentiment analysis engine trained on retail investor language, alerting infrastructure, and a historical tracking database.

Operational Cost

The honest comparison: most firms don't monitor boards at all because the manual effort doesn't scale. The cost of not monitoring is invisible — until it isn't.

What Wasn't Possible Before

Daily systematic monitoring across multiple boards for every client. No human can read hundreds of bulletin board posts daily across 15 clients and filter signal from noise. This capability simply didn't exist before AI made it economically viable.

Monthly or quarterly client reports that currently take hours to compile — coverage summary and sentiment trends, peer comparison, share price and volume context, key metrics and activity log.

The AI pulls data from the platform, generates a draft report in Alma's template, and a human reviews before sending. Time saved: 2–4 hours per client per reporting period.

Cost Saving

£18,000–72,000/year in report compilation time, reduced to £2,640/year in tool costs.

The numbers:

• 15 clients × 2–4 hours per report × 12 months = 360–720 hours/year
• At £50–100/hour blended cost = £18,000–72,000/year freed
• Annual saving: £15,000–69,000 vs £2,640 tool cost

Competitive Landscape

Many IR firms struggle with reporting timeliness — reports are delayed because they compete with more urgent client work. Automated first drafts remove the bottleneck, ensuring every client gets a timely, comprehensive report without straining consultant capacity.

Build Cost

Pre-AI estimate: £40,000–80,000 build + ongoing template maintenance

Building equivalent capability without AI would mean developing a data aggregation pipeline, report template engine, and automated chart/table generation.

Operational Cost

What Wasn't Possible Before

Comprehensive data aggregation across all platform tools in a single report. Manual reports are limited by whatever the consultant has time to pull together. The AI draws from every tool's data — coverage, sentiment, share price, peer comparisons — producing a more complete picture than any manual compilation could achieve in the same timeframe.

Before any client meeting or call, get a comprehensive briefing pack: recent coverage summary, share price and trading context, recent announcements and market reaction, anticipated questions based on current situation, and relevant peer activity.

Consultant inputs 'meeting with [Client] CEO tomorrow' and gets a structured briefing. Time saved: 30–60 minutes per significant meeting.

Cost Saving

£11,400–45,600/year in meeting preparation time, reduced to £600/year in tool costs.

The numbers:

• 38 meetings/month × 30–60 min each = 19–38 hours/month
• At £50–100/hour blended cost = £950–3,800/month freed
• Annual saving: £11,000–45,000 vs £600 tool cost

Competitive Landscape

Most firms handle meeting prep through institutional knowledge and experience. A structured, data-driven briefing pack means even junior consultants walk into meetings with senior-level preparation — levelling up the entire team without depending on individual knowledge.

Build Cost

Pre-AI estimate: £30,000–60,000 build

Building equivalent capability without AI would mean developing a data aggregation layer pulling from multiple sources, a context engine, and a briefing template system.

Operational Cost

What Wasn't Possible Before

Comprehensive briefings for every meeting, not just important ones. Manual prep means consultants triage — thorough preparation for board meetings, rushed or skipped for routine calls. The AI generates a full briefing with peer context and anticipated questions in 2 minutes, making every client interaction equally well-prepared.

After any significant announcement, automatically generate an impact report showing share price movement (immediate, 1-day, 5-day), trading volume vs average, how it compared to peer announcements, media and bulletin board sentiment, and a narrative summary of market reception.

This lets Alma demonstrate value to clients with concrete data, learn what works through pattern recognition, and use the analysis as a new business tool — 'here's how your announcements have landed vs peers.'

Revenue Opportunity

Announcement Impact Analysis doubles as a sellable deliverable and a new business development tool.

The numbers:

• Alma's cost per analysis: ~£0.06 (AI costs)
• Charge to clients: £1,000–3,000 per analysis
• At 2 analyses/client/year (15 clients): £30,000–90,000/year in additional revenue
• As a retainer add-on: £500–1,000/quarter per client = £30,000–60,000/year

New Business Weapon

Run a competitor's announcement impact analysis before a pitch. Show a prospect exactly where their communications are underperforming relative to peers — with data, not opinion. This positions Alma as an evidence-based advisory firm.

Competitive Landscape

Most financial PR firms provide anecdotal post-announcement feedback ("the market reacted well"). Structured, data-driven impact reports with peer benchmarking are genuinely rare. As the analysis history grows, Alma can identify what communication strategies consistently drive better market reception — turning accumulated data into strategic insight.

Build Cost

Pre-AI estimate: £100,000–180,000 build + £15,000–25,000/year maintenance

Building equivalent capability without AI would mean developing integrations with share price APIs, trading volume data, media monitoring, bulletin board scrapers, and a statistical comparison engine.

Operational Cost

In practice, most announcements go without formal analysis because the manual effort is too expensive. Firms settle for a quick share price check the next morning.

What Wasn't Possible Before

Historical pattern analysis across all past announcements. Manually tracking how different communication strategies affect market reception over dozens of announcements is impractical. The AI builds a cumulative picture — identifying what works and what doesn't — something that would require a dedicated research analyst to maintain.

Data-driven investor targeting to help Alma advise clients on which investors to approach for roadshows, placings, or ongoing IR outreach.

Starting with a light version using free data (Companies House, RNS major holdings disclosures) for peer shareholder analysis — identifying funds that own peers but not the client, and holders that are underweight vs comparable positions.

Designed to upgrade to mid-tier data sources if demand warrants, increasing capability significantly.

Cost Saving

£36,000–144,000/year in analyst research time, reduced to £2,400/year in tool costs. Or compared to premium data subscriptions: £75,000–150,000/year across 15 clients for Argus Vickers, vs £2,400/year.

The numbers:

• 15 analyses/month × 4–8 hours each = 60–120 hours/month
• At £50–100/hour blended cost = £3,000–12,000/month freed
• Annual saving: £34,000–142,000 vs £2,400 tool cost

Client Value

For companies preparing for roadshows, placings, or capital raises, data-driven investor targeting is directly tied to fundraising success. "These 15 funds own your 4 closest peers but not you" is materially more compelling than a generic investor list from a broker.

Competitive Landscape

Most small and mid-cap IR firms rely on their broker network. Larger firms use expensive data services. This tool provides 70–80% of the insight at a fraction of the cost, with an upgrade path to premium data if clients value deeper analysis.

Build Cost

Pre-AI estimate: £80,000–150,000 build + £10,000–20,000/year maintenance

Building equivalent capability without AI would mean developing Companies House and RNS data integrations, a shareholder cross-referencing engine, fund profile matching, and a scoring/ranking system.

Alternatively, premium data services (Argus Vickers, Ipreo) cost £5,000–10,000/year per client — and still require analyst time to interpret.

Operational Cost

What Wasn't Possible Before

Automated peer shareholder cross-referencing. Manually comparing shareholder registers across 5 peers to identify funds that own competitors but not the client is a multi-day research project. The AI does this in minutes, across every peer, with underweight position detection — a level of analysis that was only available to firms paying for expensive institutional data services.